Dissertations / Theses on the topic 'Lexical feature'

Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004.<br>Includes bibliographical references (leaves 150-151).<br>The focus of this thesis was the design, implementation, and evaluation of a set of automated algorithms to detect nasal consonants from the speech waveform in a distinctive feature-based speech recognition system. The study used a VCV database of over 450 utterances recorded from three speakers, two male and one female. The first stage of processing for each speech waveform included automated 'pivot' estimation using the Consonant Landmark Detector - these 'pivots' were considered possible sonorant closures and releases in further analyses. Estimated pivots were analyzed acoustically for the nasal murmur and vowel-nasal boundary characteristics. For nasal murmur, the analyzed cues included observing the presence of a low frequency resonance in the short-time spectra, stability in the signal energy, and characteristic spectral tilt. The acoustic cues for the nasal boundary measured the change in the energy of the first harmonic and the net energy change of the 0-350Hz and 350-1000Hz frequency bands around the pivot time. The results of the acoustic analyses were translated into a simple set of general acoustic criteria that detected 98% of true nasal pivots. The high detection rate was partially offset by a relatively large number of false positives - 16% of all non-nasal pivots were also detected as showing characteristics of the nasal murmur and nasal boundary. The advantage of the presented algorithms is in their consistency and accuracy across users and contexts, and unlimited applicability to spontaneous speech.<br>by Neira Hajro.<br>M.Eng.

<p> This study investigated two classifications of semantic features, feature importance and feature relevance, to verify if they differentially influence lexical semantic knowledge in individuals with aphasia. Feature <i>importance </i> is defined as "how important a feature is in defining a concept" (Hampton,1979), while feature relevance represents the "core meaning of a concept" (Sartori, Lombardi &amp; Mattiuzi, 2005). </p><p> A sorting task was used with 20 volunteer participants with aphasia to investigate the semantic processing involved in the association of semantic features with nouns. A corpus of 18 nouns was displayed in front of each participant in groups of three along with a card containing the word "UNRELATED." The participants were given a deck of 18 cards containing features corresponding to the nouns and to the unrelated category, and were verbally instructed to sort the deck of cards into each of the four designated piles. The semantic features on the cards were rated as high, mid and low importance (HI, MI, LI) and <i>high, mid</i> and <i>low</i> relevance (HR, MR, LR). </p><p> Analysis was completed using a two-way between-subjects ANOVA to determine was whether the mean scores at the three different levels (e.g., low, mid and high) of importance and relevance differed, and to analyze if there was an interaction between the two classifications. The participants were able to assign <i>high importance</i> features with nouns more accurately than they did <i>mid</i> and <i>low importance</i> features. Feature <i>relevance</i> did not differentially influence noun-feature association. These results indicated that the ability of individuals with aphasia to accurately associate features with nouns is influenced by levels of feature importance. </p><p> In conclusion, this study found that individuals with aphasia are more cognitively sensitive to <i>high</i> level versus <i>low</i> level feature importance and the effect does not extend to a<i> mid</i> level of importance. The study also demonstrated that the levels of feature <i> relevance</i> did not differentially influence the ability of individuals with aphasia to associate semantic features with their appropriate nouns. Potential clinical implications and study limitations were discussed.</p>

Lexical access is the process in which basic components of meaning in language, the lexical entries (words) are activated. This activation is based on the organization and representational structure of the lexical entries. Semantic features of words, which are the prominent semantic characteristics of a word concept, provide important information because they mediate semantic access to words. An experiment was conducted to examine the importance of semantic feature distinctiveness and feature frequency in accessing the lexical representations of young and older adults in an off-line task using features of animals. The McRae, Cree, Seidenberg, and McNorgan (2005) feature norm corpus is the basis for the selection of stimuli for the current research project. Semantic features were utilized to explore the structure of the lexicon. Stimuli varied in feature distinctiveness based on the study by McRae, et al. (2005) in 3 broad stimulus groups: Distinctive (D), Low Frequency Non-Distinctive (LFND), and Non-Distinctive High Frequency (NDHF). Participants were asked to list all of the concepts that came to mind for a given feature in an untimed task. Distinctiveness was examined between stimulus groups for the number of concepts and variety of first concepts given to the presented feature. It was found that fewer concepts were given and there was less variety in first concepts given for the distinctive features and the most concepts and greater variety of first concepts were given for the high-frequency non-distinctive features. Distinctiveness appears to vary along a continuum, supporting theories of lexical access based on activation and competition between concept words. Additionally, participant age groups were compared for the number of concepts given and the variety of first concepts given. The older adult group produced more concepts and more variety of first concepts than the younger group, in all three feature categories. These results indicate that greater (lifetime) language experience of the participants in the older group was reflected in their performance. A continued interest in semantic features is important to our understanding of the influence of features on the retrieval of semantic concepts and the changes in those retrieval processes over the lifespan.

UAM CorpusTool is a set of tools for the linguistic annotation of text. Core concepts include: • The user defines a ‘project’, which is a set of files, and a set of analyses which are applied to each of these files. • Each ‘analysis’ can be seen as a ‘layer’ of annotation. CorpusTool currently allows two types of annotation: 1. Document Coding: where the text as a whole is assigned features. For instance, these features could represent the register of the document (field,tenor, mode), or text-type. 2. Segment Coding: The user can select segments within a file, and assign features to each of these segments. Segments are specified by dragging the mouse over a span of text, and the user is then prompted to specify the features of this segment. Other annotation types will be added in later versions, allowing annotation of rhetorical structure theory (RST), Generic Structure (GSP), participant chaining, sentence structuring (e.g., Subj, Pred, Mood, Adjunct, etc.), annotation of spoken data etc. UAM CorpusTool replaces prior software of the author, Systemic Coder, which allowed coding of single documents at a single layer. CorpusTool is an attempt to overcome the various limitations that constrained users of Coder. I wish to thank the many users of Coder who forwarded their comments over the years, and to thank those sending me comments on this new tool.

Automatic emotion recognition has long been a focus of Affective Computing. It has become increasingly apparent that awareness of human emotions in Human-Computer Interaction (HCI) is crucial for advancing related technologies, such as dialogue systems. However, performance of current automatic emotion recognition is disappointing compared to human performance. Current research on emotion recognition in spoken dialogue focuses on identifying better feature representations and recognition models from a data-driven point of view. The goal of this thesis is to explore how incorporating prior knowledge of human emotion recognition in the automatic model can improve state-of-the-art performance of automatic emotion recognition in spoken dialogue. Specifically, we study this by proposing knowledge-inspired features representing occurrences of disfluency and non-verbal vocalisation in speech, and by building a multimodal recognition model that combines acoustic and lexical features in a knowledge-inspired hierarchical structure. In our study, emotions are represented with the Arousal, Expectancy, Power, and Valence emotion dimensions. We build unimodal and multimodal emotion recognition models to study the proposed features and modelling approach, and perform emotion recognition on both spontaneous and acted dialogue. Psycholinguistic studies have suggested that DISfluency and Non-verbal Vocalisation (DIS-NV) in dialogue is related to emotions. However, these affective cues in spoken dialogue are overlooked by current automatic emotion recognition research. Thus, we propose features for recognizing emotions in spoken dialogue which describe five types of DIS-NV in utterances, namely filled pause, filler, stutter, laughter, and audible breath. Our experiments show that this small set of features is predictive of emotions. Our DIS-NV features achieve better performance than benchmark acoustic and lexical features for recognizing all emotion dimensions in spontaneous dialogue. Consistent with Psycholinguistic studies, the DIS-NV features are especially predictive of the Expectancy dimension of emotion, which relates to speaker uncertainty. Our study illustrates the relationship between DIS-NVs and emotions in dialogue, which contributes to Psycholinguistic understanding of them as well. Note that our DIS-NV features are based on manual annotations, yet our long-term goal is to apply our emotion recognition model to HCI systems. Thus, we conduct preliminary experiments on automatic detection of DIS-NVs, and on using automatically detected DIS-NV features for emotion recognition. Our results show that DIS-NVs can be automatically detected from speech with stable accuracy, and auto-detected DIS-NV features remain predictive of emotions in spontaneous dialogue. This suggests that our emotion recognition model can be applied to a fully automatic system in the future, and holds the potential to improve the quality of emotional interaction in current HCI systems. To study the robustness of the DIS-NV features, we conduct cross-corpora experiments on both spontaneous and acted dialogue. We identify how dialogue type influences the performance of DIS-NV features and emotion recognition models. DIS-NVs contain additional information beyond acoustic characteristics or lexical contents. Thus, we study the gain of modality fusion for emotion recognition with the DIS-NV features. Previous work combines different feature sets by fusing modalities at the same level using two types of fusion strategies: Feature-Level (FL) fusion, which concatenates feature sets before recognition; and Decision-Level (DL) fusion, which makes the final decision based on outputs of all unimodal models. However, features from different modalities may describe data at different time scales or levels of abstraction. Moreover, Cognitive Science research indicates that when perceiving emotions, humans make use of information from different modalities at different cognitive levels and time steps. Therefore, we propose a HierarchicaL (HL) fusion strategy for multimodal emotion recognition, which incorporates features that describe data at a longer time interval or which are more abstract at higher levels of its knowledge-inspired hierarchy. Compared to FL and DL fusion, HL fusion incorporates both inter- and intra-modality differences. Our experiments show that HL fusion consistently outperforms FL and DL fusion on multimodal emotion recognition in both spontaneous and acted dialogue. The HL model combining our DIS-NV features with benchmark acoustic and lexical features improves current performance of multimodal emotion recognition in spoken dialogue. To study how other emotion-related tasks of spoken dialogue can benefit from the proposed approaches, we apply the DIS-NV features and the HL fusion strategy to recognize movie-induced emotions. Our experiments show that although designed for recognizing emotions in spoken dialogue, DIS-NV features and HL fusion remain effective for recognizing movie-induced emotions. This suggests that other emotion-related tasks can also benefit from the proposed features and model structure.

This thesis investigates Swedish lexical blends. A lexical blend is defined as the concatenation of two words, where at least one word has been reduced. Lexical blends are approached from two perspectives. First, the thesis investigates lexical blends as they appear in the Swedish language. It is found that there is a significant statistical relationship between the two source words in terms of orthographic, phonemic and syllabic length and frequency in a reference corpus. Furthermore, some uncommon lexical blends created from pronouns and interjections are described. A description of lexical blends through semantic construction and similarity to other word formation processes are also described. Secondly, the thesis develops a model which predicts source words of lexical blends. To predict the source words a logistic regression model is used. The evaluation shows that using a ranking approach, the correct source words are the highest ranking word pair in 32.2% of the cases. In the top 10 ranking word pairs, the correct word pair is found in 60.6% of the cases. The results are lower than in previous studies, but the number of blends used is also smaller. It is shown that lexical blends which overlap are easier to predict than lexical blends which do not overlap. Using feature ablation, it is shown that semantic and frequency related features have the most important for the prediction of source words.

Il presente lavoro di ricerca intende indagare le caratteristiche del linguaggio rivolto ai bambini nella lingua italiana nel primo anno di vita e i suoi effetti sullo sviluppo del linguaggio, dai prerequisiti allo sviluppo lessicale e sintattico. La cornice teorica su cui si fondano le ricerche qui presentate assume che il processo di acquisizione linguistica abbia basi sociali. La prima parte dell’elaborato comprende due studi che descrivono le proprietà lessicali, sintattiche e prosodiche del linguaggio rivolto ai bambini. La seconda, invece, è costituita da due lavori che indagano la qualità e gli effetti dell’input linguistico nello sviluppo del linguaggio, prendendo in considerazione sia un prerequisito in fase verbale, sia le competenze lessicali e sintattiche nel secondo anno di vita; in questa seconda parte sono stati inoltre definiti i predittori dell’apprendimento linguistico, considerando sia le caratteristiche dell’input ma anche il contributo delle competenze comunicative precoci del bambino. Il primo studio presentato è un’indagine a carattere longitudinale nella quale sono state descritte, mediante misure globali e specifiche, le caratteristiche lessicali e sintattiche del linguaggio rivolto ai bambini nella lingua italiana. Ciò che è emerso è un registro semplificato ma non semplice che presenta un periodo di massima semplificazione nella seconda metà del primo anno di vita. La seconda ricerca, sempre a carattere longitudinale, ha preso in esame le proprietà prosodiche del linguaggio rivolto ai bambini e la caratterizzazione prosodica di enunciati con funzione pragmatica differente. I risultati hanno messo in luce una prosodia generalmente enfatizzata nel linguaggio rivolgo ai bambini nel periodo preverbale ma, sorprendentemente, in misura moderata. Inoltre, è stato possibile osservare un pattern di cambiamento nel corso del primo anno di vita che si discosta da quello caratterizzante altre lingue non-tonali. Infine, sono emerse caratteristiche prosodiche distintive per enunciati con funzione pragmatica diversa, elemento che evidenzia il ruolo altamente informativo della prosodia. Il terzo lavoro ha indagato longitudinalmente gli antecedenti dello sviluppo linguistico, valutando il contributo delle competenze comunicative precoci del bambino e il ruolo dell’input - di cui sono state esaminate qualità e stabilità - temi rispetto ai quali la letteratura riporta ancora risultati contrastanti. I dati ottenuti indicano che lo sviluppo linguistico nel secondo anno di vita rispecchi le abilità comunicative precoci e sembri favorito da un input ricco, ridondante e sintatticamente articolato. Infine, il quarto contributo ha analizzato, con un disegno sperimentale, i possibili effetti del canto rivolto ai bambini, ipotizzando un ruolo facilitatore rispetto al parlato nel processo di discriminazione fonetica, precursore preverbale dello sviluppo linguistico. Si tratta di un tema piuttosto trascurato nella letteratura che, invece, si è di fatto sempre concentrata sugli effetti della prosodia tipica del parlato rivolto ai bambini. I risultati principali hanno messo in luce come il ruolo facilitatore del canto in tale registro emerga alla fine del primo anno di vita quando, da un punto di vista evolutivo, si verifica un cambiamento nella capacità di discriminare i fonemi nativi e non nativi. È stato altresì possibile individuare benefici di una maggiore esposizione alla musica e al canto in fase preverbale, sia rispetto alla discriminazione fonetica che al successivo sviluppo lessicale.<br>This research work aims to explore infant-directed speech features in Italian language during the first year of an infant’s life and its effects on language acquisition, from precursors to advanced lexical and syntactic skills. The theoretical background assumes social bases of linguistic development. The first part consists of two studies on lexical, syntactic and prosodic properties in this special register. The second part includes two researches on quality and effects of linguistic input in language acquisition, taking into account a preverbal precursor and lexical and syntactic abilities during the second year of life; additionally, in this section, the predictors of language learning have been defined, exploring the role of linguistic input and the contribution of early communication skills in infants. The first study is a longitudinal design investigation, with an exhaustive analysis of lexical and syntactic characteristics of infant-directed speech in Italian language, comprehensive of both global and specific measures. From this investigation, the special register addressed to infants appears as a simplified but not simple with a period of maximum simplification in the second half of the first year of an infant’s life. The second longitudinal research examines prosodic properties in infant-directed speech and prosodic characterization of utterances with different pragmatic function. Results show how typical prosody in Italian infant-directed speech is overall emphasized in the preverbal period but, surprisingly, moderately; moreover, prosody changes during the first year even though without the same pattern of other non-tonal languages. Lastly, utterances with different pragmatic functions are characterized by a distinctive prosody. In the third contribution, predictors of language acquisition are longitudinally explored, analyzing the role of early communication skills in infants and of maternal input. In addition, input quality and stability are evaluated. About this topic, literature shows conflicting results. Overall, we find how subsequent linguistic abilities could be predicted by infant’s early communication skills and a by a rich, redundant, syntactically articulated but lexically repetitive input. Lastly, the fourth experimental work analyses the facilitator role of infant-directed song compared to infant-directed speech on the phonetic discrimination process, a preverbal precursor of language acquisition. Literature highlights how typical prosody in this special speech supports the identification of linguistic units in the verbal flow. Nevertheless, the role of infant-directed song has been poorly explored, especially as regard the development of a linguistic prerequisite. Main results prove a facilitator role of infant-directed song at the end of the first year of an infant’s life, when changes in the phonetic discrimination skill occur. Moreover, we find benefic effects of an higher musical and song exposition during the preverbal stage on both phonetic discrimination and subsequent lexical skills.

The purpose of this study was to explore stuttering behaviour associated with word-type and word-frequency in a large group of adults who stutter (AWS). The study method involved analysing spontaneous speech samples of 50 AWS on the following features: 1) the occurrence of stuttering on content and function words, 2) the occurrence of stuttering on function words at the function/content word boundary, 3) the occurrence of syllable repetitions in relation to moments of stuttering on function and content words, and 4) word frequency of stuttered words compared to control words. Results indicated that moments of stuttering occurred more often on content words compared to function words. Furthermore, if stuttering occurred on a function word, it was found to occur most often at the function/content word boundary. No difference was found when the data were further analyzed as a function of frequency of stuttering. The AWS who had a high frequency of stuttering did not differ significantly from those with a low frequency of stuttering. Syllable repetitions occurred more often on function words compared to content words. In general, the AWS who showed a low frequency of stuttering exhibited a higher percentage of syllable repetitions. The word frequency of stuttered versus control words did not differ significantly. There was also no difference in word frequency for stuttered function versus stuttered content words. The results were interpreted to provide general support for the EXPLAN theory of stuttering behaviour proposed by Howell (2004) and colleagues (Howell & Au Yeung, 2002; Howell, Au-Yeung & Sackin, 2000).

L’onomatopée et l’interjection qui ont été étudiées de manière approfondie dans le domaine français dès le XVIIIe siècle n’ont fait l’objet que de quelques allusions dans la théorie grammaticale arabe traditionnelle. C’est ce vide que la présente thèse se propose de combler. Il s’agit d’une étude lexicale qui s’insère dans les recherches développant la théorie des matrices et des étymons (TME). Elle comprend deux grandes parties. La première se compose de deux chapitres. Le premier est consacré aux interjections (34 cas) et le second aux onomatopées proprement dites (18 cas). Dans ces deux chapitres, on procède à la manière de Guiraud (1967, chapitre III, « Structures onomatopéiques », ex. la racine T. K.). On extrait de l’interjection ou de l’onomatopée les deux consonnes qui les composent et on les considère comme un étymon dont on suit les diverses réalisations dans le lexique, en veillant toujours à la corrélation phono-sémantique. Il ressort de cette première partie que les racines (au sens traditionnel du terme) émanant d’étymons onomatopéiques sont près de neuf cents.Comme nous avons observé la présence du trait [+strident] dans un bon nombre de ces étymons, nous avons consacré la seconde partie à l’étude d’une nouvelle matrice qui se compose des traits [+strident] et [coronal], corrélés l’invariant notionnel « son, voix, bruit » et ses conséquences ,incluant la peur, le mouvement et ses diverses caractérisations. Cette matrice se réalise dans 69 étymons qui sont disposés par ordre alphabétique. Les réalisations de chaque étymon sont disposées en trois colonnes : le radical où l’étymon se réalise, la référence dans le tableau de l’organisation de l’invariant notionnel, le sens (tel qu’il est donné dans le Kazimirski).Il ressort de cette double étude qu’un immense pan du lexique de l’arabe est motivé, pourvu que l’on situe l’analyse au niveau des traits, la corrélation s’établissant entre le bruit et la définition même du trait de stridence, les sons stridents « se caractérisant acoustiquement par une plus grande quantité de bruit que leurs correspondants non-stridents » (Chomsky et Halle, 1968, p. 166)<br>Although onomatopoeia and interjections in French have been studied in depth since the 18th century, they have only been alluded to in traditional Arabic grammatical theory. This thesis is intended to remedy this situation with a lexical study applying the Theory of Matrices and Etymons (TME). The thesis is divided into two major sections, the first of which comprises two chapters. The first chapter deals with 34 interjections and the second 18 onomatopoeia. The approach adopted in these first two chapters is that of Guiraud (1967, chapter III devoted to onomatopoeic structures, e.g. the T. K. root). The interjections and onomatopoeia are extracted from the consonants which compose them; these consonants are considered as an etymon the various realisations of which are pursued in the lexicon with constant attention to the phono-semantic correlation. This first part shows that there are nearly nine hundred roots (in the traditional meaning of the term) arising out of the onomatopoeic etymons. Given the presence of the [+strident] feature in a large number of these etymons, the second part of this thesis is devoted to a new matrix comprising the correlated [+strident] and [coronal] features, the notional invariant "sound, voice, noise" and its consequences including fear, movement and its various characteristics. This matrix is realised in 69 etymons ordered alphabetically. The realisations of each etymon are given in three columns: the radical in which the etymon is realised; the reference in the organisational table of the notional invariant; the meaning (as given in Kazimirski).This double study reveals that - provided that the analysis is established at the phonetic feature level - an immense area of the Arabic lexicon is motivated, the correlation between the noise and the very definition of the feature of stridence and strident sounds: "Strident sounds are marked acoustically by greater noisiness than their nonstrident counterparts" (Chomsky et Halle, 1968, p. 329)

The main results of the PhD thesis are described in the summary. The main objective of the thesis is to define the acquisition of lexical and morphosyntactic features of Lithuanian adjective. The analysis is based on the large quantity of linguistic data (~ 400 000 words), which is the corpus of four children compiled according the method of longitudinal observation. The study has interchangeably applied several methods of analysis: longitudinal observation, corpus linguistics, error analysis and comparative method. Statements to be Defended 1. Lexical features of adjectives are acquired by children with difficulties: they often confuse adjectives of a single semantic group and make many errors. 2. Children easily acquire morphosyntactic features of adjectives: they do not make many errors. Some errors may be observed only in cases, when adjective paradigms do not coincide with noun paradigms. 3. Lexical and morphosyntactic features of adjectives are acquired simultaneously. The meaning of adjective is realised only when multi-word combinations and morphosyntactic features have been acquired. 4. The acquisition of adjectives correlates with the individual language development. Bigger lexical variety of adjectives is more characteristic to children with early and late language development, if compared to children with typical language development. Morphosyntactic features of adjectives do not pose problems to children with late language development. 5. The acquisition of... [to full text]<br>Santraukoje pateikiami pagrindiniai disertacinio tyrimo metu gauti rezultatai. Aprašomo tyrimo tikslas – išanalizuoti leksines ir morfosintaksines lietuvių kalbos būdvardžio įsisavinimo ypatybes. Tyrimui pasirinktas didelės apimties šaltinis – keturių vaikų kalbos tekstynas (~ 400 000 žodžių), sukauptas ilgalaikio stebėjimo metodu. Tyrimas atliktas taikant ir derinant kelis metodus: ilgalaikio stebėjimo, tekstynų lingvistikos, klaidų analizės ir lyginamąjį. Atlikus tyrimą darbe pagrindžiami šie ginamieji teiginiai: 1. Vaikams sudėtinga įsisavinti būdvardžio leksines ypatybes: dažnai painiojami vienos leksinės semantinės grupės būdvardžiai, daroma daug klaidų. 2. Vaikai lengvai įsisavina morfosintaksines būdvardžio ypatybes: klaidų nėra daug, dažniau klystama tik retų paradigmų būdvardžius derinant su daiktavardžiais. 3. Būdvardžio leksinės ir morfosintaksinės ypatybės įsisavinamos drauge. Būdvardžio reikšmė suvokiama tik pradėjus vartoti keliažodžius pasakymus, įsisavinus gramatines ypatybes. 4. Būdvardžio įsisavinimas susijęs su individualia vaiko kalbos raida. Ankstyvajai ir vėlyvajai kalbos raidai būdinga didesnė būdvardžių leksinė įvairovė nei įprastai kalbos raidai, o vaikams, kurių kalbos raida vėlyva, nesudėtinga įsisavinti būdvardžio morfosintaksinius požymius. 5. Būdvardžio įsisavinimą veikia ne tik aplinkos veiksniai (vaikiškoji kalba), bet ir kalbos sistema. Kaip ir įsisavinant kitas kalbos dalis, vaikų kalboje vyrauja vaikiškojoje kalboje dažniausiai vartojamos... [toliau žr. visą tekstą]

The thesis discusses the variety of English that is spoken today in Brunei Darussalam and assesses its status as a 'New English'. Using a corpus of spoken data which was recorded and transcribed by the author, the thesis attempts to produce an empirically based linguistic description of the grammatical, lexical and discourse features found in spoken Brunei English and to discuss the ways in which these features differ from the equivalent features in Standard British English. The final part of the study is concerned with the pedagogical and language planning implications of recognizing the existence of a Bruneian variety of English, and with proposing an appropriate English language teaching model for the Bruneian education system.

Cette thèse établit un cadre de travail de reconnaissance de formes pour les systèmes de reconnaissance de mots hors-ligne. Elle se concentre sur les caractéristiques de l’image, car elles ont une grande influence sur les performances de reconnaissance. En particulier, nous considérons deux aspects complémentaires de l’impact des caractéristiques: la réduction du lexique et la reconnaissance elle-même. Le premier aspect, la réduction du lexique, consiste à concevoir un classifieur faible qui fournit en sortie un ensemble d’hypothèses de mots à partir d’une image de mot. Son objectif principal est de réduire le temps de calcul de la reconnaissance tout en maintenant (voire améliorant) le taux de reconnaissance. Le deuxième aspect est le système de reconnaissance proprement dit. Plusieurs caractéristiques existent dans la littérature, issues de différents domaines de recherche, mais il n’existe pas de consensus sur les pistes les plus prometteuses. L’objectif de cette thèse est d’améliorer notre compréhension des caractéristiques pertinentes pour construire des systèmes de reconnaissance encore plus performants. À cette fin, nous avons abordé deux problèmes spécifiques: 1) la conception de caractéristiques pour la réduction du lexique (appliquée à l’écriture arabe), et 2) l’évaluation de caractéristiques pour la reconnaissance de l’écriture manuscrite cursive (appliquée à l’écriture latine et arabe). Contrairement à l’écriture latine, la problématique de réduction du lexique est peu abordée pour l’écriture arabe. Les méthodes existantes utilisent certaines caractéristiques fondamentales des mots arabes telles que le nombre de sous-mots et les signes diacritiques, mais ignorent totalement la forme des sous-mots. Par conséquent, notre premier objectif est de créer une méthode de réduction du lexique basée sur la forme des sous-mots. Notre approche utilise l’indexation de formes, où la forme d’un sous-mot requête est comparée à une base de données étiquetée de sous-mots échantillons. Pour une comparaison efficace avec un faible temps de calcul, nous avons proposé le vecteur de signature topologique pondéré (W-TSV), où la forme du sous-mot est modélisée par un graphe acyclique orienté (DAG) pondéré, à partir duquel le vecteur W-TSV est extrait pour l’indexation. La principale contribution de ce travail est d’élargir le cadre existant du vecteur de signature topologique (TSV) aux DAGs pondérés et de proposer une approche d’indexation de formes pour la réduction du lexique. Cette approche est performante pour la réduction d’un lexique composé de sous-mots arabes. Néanmoins, ses performances restent modestes pour les mots arabes. Compte tenu des résultats de notre premier travail sur la réduction du lexique de mots arabe, nous proposons de construire un nouvel index pour de meilleures performances au niveau du mot. La forme des sous-mots, ainsi que leur nombre et celui des signes diacritiques sont des éléments importants de la forme du mot arabe. Nous proposons donc le descripteur de mot arabe (AWD) qui intègre toutes les composantes mentionnées ci-dessus. Il est construit en deux étapes. Tout d’abord, un descripteur de structure (SD) est calculé pour chaque composante connexe (CC) d’une image de mots. Il décrit la forme de la CC en utilisant le modèle de sac-de-mots, où chaque mot visuel représente une structure locale particulière. Ensuite, l’AWD est formé par la concaténation des SDs en utilisant une heuristique efficace, qui différencie implicitement les sous-mots des signes diacritiques. Dans le contexte de la réduction du lexique, l’AWD est utilisé pour indexer une base de données référence. La principale contribution de ce travail est la conception de l’AWD, qui intègre les caractéristiques de bas niveau (structure de la forme du sous-mot) et les informations symboliques (nombre de sous-mots et de signes diacritiques) en un seul descripteur. La méthode proposée possède un faible temps de calcul et elle est facile à implémenter. Elle fournit de meilleures performances pour la réduction du lexique sur deux bases de données d’écriture arabe, à savoir la base de données de sous-mots Ibn Sina et la base de données de mots IFN/ENIT. La dernière partie de cette thèse se concentre sur les caractéristiques visuelles pour la reconnaissance de mots. Un grand nombre de caractéristiques existent dans la littérature, chacune d’elles étant motivées par différents domaines, tels que la reconnaissance des formes, la vision par ordinateur ou l’apprentissage machine. Identifier les approches les plus prometteuses servirait à améliorer la conception de la prochaine génération de caractéristiques. Néanmoins, comme elles sont fondées sur des concepts différents, il est difficile de les comparer de manière théorique, des outils empiriques sont donc nécessaires. Par conséquent, le dernier objectif de la thèse est de fournir une méthode d’évaluation de caractéristiques en fonction de leur force et complémentarité. Un modèle de combinaison a été conçu à cet effet, dans lequel chaque caractéristique est évaluée au travers d’un système référence de reconnaissance, basée sur les réseaux de neurones récurrents. Plus précisément, chaque caractéristique est représentée par un agent, qui est une instance du système de reconnaissance entraînée à partir de cette caractéristique. Les décisions de tous les agents sont combinées en utilisant un vote pondéré. Les poids sont optimisés conjointement au cours d’une phase d’entraînement, afin d’augmenter le vote pondéré de la véritable étiquette de chaque mot. Par conséquent, les poids reflètent la force et la complémentarité des agents et de leurs caractéristiques pour la tâche donnée. Enfin, les poids sont convertis en scores numériques attribués aux caractéristiques, qui sont faciles à interpréter sous ce modèle de combinaison. Au meilleur de notre connaissance, c’est la première méthode d’évaluation de caractéristiques capable de quantifier l’importance de chaque caractéristique, au lieu d’établir un classement basé sur le taux de reconnaissance. Cinq caractéristiques de l’état de l’art ont été testées et nos résultats offrent une perspective intéressante pour la conception de futures caractéristiques.

Unlike some Asian languages (e.g., Korean), English has lexical stress manifested by four acoustic features: duration, intensity, F0 (pitch), and vowel quality. Lexical stress has been known to have significant influences on native English speakers’ recognition of spoken words. According to Cutler (2015), lexical stress has both suprasegmental and segmental features: Suprasegmental features include duration, intensity, and F0 while vowel quality is considered a segmental feature. However, it is still unclear which lexical features are more responsible for spoken word recognition. This study examined which features, suprasegmental features or vowel quality of English, are a more significant influencer in spoken word recognition using English loanwords in Korean, which lack the prominence of any syllable realized by these features. Additionally, this study investigated the claimed advantage of the strong-weak stress pattern over a weak-strong pattern. To that end, two experiments were conducted. First, a parallel acoustic comparison was made between disyllabic English words and their corresponding English loanwords in Korean in order to investigate whether Korean has lexical stress features similar to those of English. 10 Korean and 10 English native speakers read 20 disyllabic words: the English loanwords in Korean by Korean participants and the source English words by American participants. The results showed that the differences of acoustic values between the syllables of the English words were significantly larger than those of the English loanwords. That is, the relative prominence of the stressed syllable over the unstressed syllable in English was not found in Korean. Additionally, the results indicated that Korean does not have a reduced vowel such as /ə/ in English, which is a critical feature of English vowel quality. In Experiment II, 16 English loanwords were used to create three versions of a spoken word recognition experiment, which was administered using the online survey platform, Qualtrics. Each version had a different type of manipulation: unmanipulated English loanwords, English loanwords with suprasegemental manipulation or English loanwords with vowel quality manipulation. 117 American English hearers identified the spoken words of one of the versions assigned to them; their success rates and reaction times (RT) were recorded. A binominal regression test was used for the analysis of success rates, and the Kruskal-Wallis H test for the response times. The results indicated that as far as success rates are concerned, both suprasegmental features and vowel quality play a role in recognizing spoken English words. However, when these two features were compared, vowel quality seemed to be a much stronger player. As for stress patterns, no significant differences were found in success rates across the three sets of manipulation. Moreover, this study did not find any significant difference in RTs either across the three manipulation sets or the two stress patterns. This study offered many applied implications in ESL, especially for teaching English pronunciation in Korea.

With advances in information and communication technologies over the past few years, the amount of information stored in the form of electronic text documents has been rapidly growing. Since the human abilities to effectively process and analyze large amounts of information are limited, there is an increasing demand for tools enabling to automatically analyze these documents and benefit from their emotional content. These kinds of systems have extensive applications. The purpose of this work is to design and implement a system for identifying expression of emotions in Czech texts. The proposed system is based mainly on machine learning methods and therefore design and creation of a training set is described as well. The training set is eventually utilized to create a model of classifier using the SVM. For the purpose of improving classification results, additional components were integrated into the system, such as lexical database, lemmatizer or derived keyword dictionary. The thesis also presents results of text documents classification into defined emotion classes and evaluates various approaches to categorization.

Informe de Seminario para optar al grado de Licenciado(a) en Lengua y Literatura Inglesa<br>This research report presents a study on vocabulary use in the L2 writing performance of a group of students from the program of English Linguistics and Literature at the Universidad de Chile. The data for the study correspond to ten tests used in a previous study (Aranda et al, 2013) and categorized in Top and Bottom groups by four teachers in relation to the quality of their performance. In addition, the study used data from four interviews to the same four teachers in order to know the specific aspects they consider when assessing vocabulary use in L2 writing performance. The tests were analyzed qualitatively in search for lexical diversity and lexical sophistication features that could account for the difference in perceived quality between the two groups. Such features included the Lexical Frequency Profile for each test, the computation of MTLD, the presence of words related to the topic of the tests and the presence of lexical errors. The interviews were transcribed and a list of the most important aspects of assessment of L2 writing according to the teachers was elaborated. The aspects identified include lexical sophistication, lexical diversity, lexical error and topic appropriateness. The results of the tests’ analyses and the interviews were contrasted, looking for the degree of coherence between them. Results show that, even when there is some degree of coherence between the tests’ analyses and what the teachers say they evaluate, lexical error is the most meaningful aspect when comparing Top and Bottom tests. This is so even when teachers considered lexical sophistication as the most important aspect for them when assessing L2 written vocabulary.

This pilot study addresses the distribution of nonstandard syntactic and lexical features in Indian English (IE) across a homogeneous group of highly educated IE speakers. It is found that nonstandard syntactic features of article use, number agreement and assignment of verb argument structure do not display uniform intragroup distribution. Instead, a relationship is found between nonstandard syntactic features and the sociolinguistic variables of lower levels of exposure to and use of English found within the group. While nonstandard syntactic features show unequal distribution, nonstandard lexical features of semantic reassignment, and mass nouns treated as count nouns display a more uniform intragroup distribution.

This mixed-methods study examines the potential existence of a military dialect separate from regional or social dialects experienced by civilians. In particular, how similar is the military-related storytelling lexicon of veterans in the Fargo-Moorhead area to the lexicon set forth in training bases and training manuals used by the U.S. military? The lexicon used by veterans in storytelling can sometimes seem opaque to an audience. It is typically dense with meaning borne by a few coded words. These words carry a contextual burden that can be better understood by an appeal to the dialect from which they were borne. In order to disentangle the veteran way of speaking from other overlapping and intersecting social and regional dialects that make up a subject’s typical speech, guided conversation and word-matching exercises were used to isolate lexicon that was typical to the military experience. The resulting interview transcripts were analyzed in comparison to military training manuals to arrive at a percentage of military-specific terms used in the guided conversation and a percentage of general knowledge military terms retained in the word-matching measure. The resulting 1.85% of military-specific terms and phrases used by participants in guided conversations and 61% retention of military-specific term knowledge was used to show that the military dialect not only exists but persists in the repertoire of veteran participants. As the majority of those who work with veterans are not veterans themselves, these percentages represent a significant barrier to understanding veteran storytelling. This barrier hinders the successful reintegration and mental health of veterans who return to their communities without knowing how to meaningfully express their stories in their existing support networks.

Sentiment analysis concerns the computational study of opinions expressed in text. Social media domains provide a wealth of opinionated data, thus, creating a greater need for sentiment analysis. Typically, sentiment lexicons that capture term-sentiment association knowledge are commonly used to develop sentiment analysis systems. However, the nature of social media content calls for analysis methods and knowledge sources that are better able to adapt to changing vocabulary. Invariably existing sentiment lexicon knowledge cannot usefully handle social media vocabulary which is typically informal and changeable yet rich in sentiment. This, in turn, has implications on the analyser's ability to effectively capture the context therein and to interpret the sentiment polarity from the lexicons. In this thesis we use SentiWordNet, a popular sentiment-rich lexicon with a substantial vocabulary coverage and explore how to adapt it for social media sentiment analysis. Firstly, the thesis identifies a set of strategies to incorporate the effect of modifiers on sentiment-bearing terms (local context). These modifiers include: contextual valence shifters, non-lexical sentiment modifiers typical in social media and discourse structures. Secondly, the thesis introduces an approach in which a domain-specific lexicon is generated using a distant supervision method and integrated with a general-purpose lexicon, using a weighted strategy, to form a hybrid (domain-adapted) lexicon. This has the dual purpose of enriching term coverage of the general purpose lexicon with non-standard but sentiment-rich terms as well as adjusting sentiment semantics of terms. Here, we identified two term-sentiment association metrics based on Term Frequency and Inverse Document Frequency that are able to outperform the state-of-the-art Point-wise Mutual Information on social media data. As distant supervision may not be readily applicable on some social media domains, we explore the cross-domain transferability of a hybrid lexicon. Thirdly, we introduce an approach for improving distant-supervised sentiment classification with knowledge from local context analysis, domain-adapted (hybrid) and emotion lexicons. Finally, we conduct a comprehensive evaluation of all identified approaches using six sentiment-rich social media datasets.

In this thesis, product experiences were analyzed in human-product interaction context. The terminology to describe product experiences were gathered and grouped with the studies carried out during the thesis. The conceptual network of product experiences were attempted to be established by examining the stages of human-product interaction in the limitations of the studies. The gathered terminology was composed of both concrete and abstract product features. The abstract product features which are used for defining product personality characteristics took part in the lexicon. The lexicon for product experience related communication resulting from the thesis, was used in designed instructional game for Bachelor degree industrial design students, in order to be informed and develop themselves on the subject. Hereby, while the students have fun, they will explore the terminology to use in design communications easily.

There is consensus among scholars that there is not just one English language but a family of “World Englishes”. The umbrella-term “World Englishes” provides a conceptual framework to accommodate the different varieties of English that have evolved as a result of the linguistic cross-fertilization attendant upon colonization, migration, trade and transplantation of the original “strain” or variety. Various theoretical models have emerged in an attempt to understand and classify the extant and emerging varieties of this global language. The hierarchically based model of English, which classifies world English as “First Language”, “Second Language” and “Foreign Language”, has been challenged by more equitably-conceived models which refer to the emerging varieties as New Englishes. The situation in a country such as multi-lingual South Africa is a complex one: there are 11 official languages, one of which is English. However the English used in South Africa (or “South African English”), is not a homogeneous variety, since its speakers include those for whom it is a first language, those for whom it is an additional language and those for whom it is a replacement language. The Indian population in South Africa are amongst the latter group, as theirs is a case where English has ousted the traditional Indian languages and become a de facto first language, which has retained strong community resonances. This study was undertaken using the methodology of corpus linguistics to initiate the creation of a repository of linguistic evidence (or corpus), of Indian South African English, a sub-variety of South African English (Mesthrie 1992b, 1996, 2002). Although small (approximately 60 000 words), and representing a narrow age band of young adults, the resulting corpus of spoken data confirmed the existence of robust features identified in prior research into the sub-variety. These features include the use of ‘y’all’ as a second person plural pronoun, the use of but in a sentence-final position, and ‘lakker’ /'lVk@/ as a pronunciation variant of ‘lekker’ (meaning ‘good’, ‘nice’ or great’). An examination of lexical frequency lists revealed examples of general South African English such as the colloquially pervasive ‘ja’, ‘bladdy’ (for bloody) and jol(ling) (for partying or enjoying oneself) together with neologisms such as ‘eish’, the latter previously associated with speakers of Black South African English. The frequency lists facilitated cross-corpora comparisons with data from the British National Corpus and the Corpus of London Teenage Language and similarities and differences were noted and discussed. The study also used discourse analysis frameworks to investigate the role of high frequency lexical items such as ‘like’ in the data. In recent times ‘like’ has emerged globally as a lexicalized discourse marker, and its appearance in the corpus of Indian South African English confirms this trend. The corpus built as part of this study is intended as the first building block towards a full corpus of Indian South African English which could serve as a standard for referencing research into the sub-variety. Ultimately, it is argued that the establishment of similar corpora of other known sub-varieties of South African English could contribute towards the creation of a truly representative large corpus of South African English and a more nuanced understanding and definition of this important variety of World English.

Disertacijoje aprašomo tyrimo tikslas – išanalizuoti leksines ir morfosintaksines lietuvių kalbos būdvardžio įsisavinimo ypatybes. Tyrimui pasirinktas didelės apimties šaltinis – keturių vaikų kalbos tekstynas (~ 400 000 žodžių), sukauptas ilgalaikio stebėjimo metodu. Tyrimas atliktas taikant ir derinant kelis metodus: ilgalaikio stebėjimo, tekstynų lingvistikos, klaidų analizės ir lyginamąjį. Atlikus tyrimą darbe pagrindžiami šie ginamieji teiginiai: 1. Vaikams sudėtinga įsisavinti būdvardžio leksines ypatybes: dažnai painiojami vienos leksinės semantinės grupės būdvardžiai, daroma daug klaidų. 2. Vaikai lengvai įsisavina morfosintaksines būdvardžio ypatybes: klaidų nėra daug, dažniau klystama tik retų paradigmų būdvardžius derinant su daiktavardžiais. 3. Būdvardžio leksinės ir morfosintaksinės ypatybės įsisavinamos drauge. Būdvardžio reikšmė suvokiama tik pradėjus vartoti keliažodžius pasakymus, įsisavinus gramatines ypatybes. 4. Būdvardžio įsisavinimas susijęs su individualia vaiko kalbos raida. Ankstyvajai ir vėlyvajai kalbos raidai būdinga didesnė būdvardžių leksinė įvairovė nei įprastai kalbos raidai, o vaikams, kurių kalbos raida vėlyva, nesudėtinga įsisavinti būdvardžio morfosintaksinius požymius. 5. Būdvardžio įsisavinimą veikia ne tik aplinkos veiksniai (vaikiškoji kalba), bet ir kalbos sistema. Kaip ir įsisavinant kitas kalbos dalis, vaikų kalboje vyrauja vaikiškojoje kalboje dažniausiai vartojamos būdvardžio leksemos ir gramatinės kategorijos. Kalbos sistema... [toliau žr. visą tekstą]<br>The main objective of the study is to define the acquisition of lexical and morphosyntactic features of Lithuanian adjective. The analysis is based on the large quantity of linguistic data (~ 400 000 words), which is the corpus of four children compiled according the method of longitudinal observation. The study has interchangeably applied several methods of analysis: longitudinal observation, corpus linguistics, error analysis and comparative method. Statements to be Defended 1. Lexical features of adjectives are acquired by children with difficulties: they often confuse adjectives of a single semantic group and make many errors. 2. Children easily acquire morphosyntactic features of adjectives: they do not make many errors. Some errors may be observed only in cases, when adjective paradigms do not coincide with noun paradigms. 3. Lexical and morphosyntactic features of adjectives are acquired simultaneously. The meaning of adjective is realised only when multi-word combinations and morphosyntactic features have been acquired. 4. The acquisition of adjectives correlates with the individual language development. Bigger lexical variety of adjectives is more characteristic to children with early and late language development, if compared to children with typical language development. Morphosyntactic features of adjectives do not pose problems to children with late language development. 5. The acquisition of adjective is influenced by the input, as well as by the linguistic... [to full text]

Machine Translation (MT) for language pairs with long distance dependencies and word reordering, such as German-English, is prone to producing output that is lexically or syntactically incoherent. Statistical MT (SMT) models used explicit or latent syntax to improve reordering, however failed at capturing other long distance dependencies. This thesis explores how explicit sentence-level syntactic information can improve translation for such complex linguistic phenomena. In particular, we work at the level of the syntactic-semantic interface with representations conveying the predicate-argument structures. These are essential to preserving semantics in translation and SMT systems have long struggled to model them. String-to-tree SMT systems use explicit target syntax to handle long-distance reordering, but make strong independence assumptions which lead to inconsistent lexical choices. To address this, we propose a Selectional Preferences feature which models the semantic affinities between target predicates and their argument fillers using the target dependency relations available in the decoder. We found that our feature is not effective in a string-to-tree system for German-English and that often the conditioning context is wrong because of mistranslated verbs. To improve verb translation, we proposed a Neural Verb Lexicon Model (NVLM) incorporating sentence-level syntactic context from the source which carries relevant semantic information for verb disambiguation. When used as an extra feature for re-ranking the output of a German-English string-to-tree system, the NVLM improved verb translation precision by up to 2.7% and recall by up to 7.4%. While the NVLM improved some aspects of translation, other syntactic and lexical inconsistencies are not being addressed by a linear combination of independent models. In contrast to SMT, neural machine translation (NMT) avoids strong independence assumptions thus generating more fluent translations and capturing some long-distance dependencies. Still, incorporating additional linguistic information can improve translation quality. We proposed a method for tightly coupling target words and syntax in the NMT decoder. To represent syntax explicitly, we used CCG supertags, which encode subcategorization information, capturing long distance dependencies and attachments. Our method improved translation quality on several difficult linguistic constructs, including prepositional phrases which are the most frequent type of predicate arguments. These improvements over a strong baseline NMT system were consistent across two language pairs: 0.9 BLEU for German-English and 1.2 BLEU for Romanian-English.

Заголовки новин допомагають зрозуміти читачеві чи потрібно продовжувати читати статтю. Вдало підібраний заголовок переконує читача зупинитися і присвятити свій час подальшому читанню. Саме тому актуальним є дослідження стилю написання та перекладу заголовків новин, які мають свої особливості, відмінні від інших типів текстів. Це стосується не тільки лексичних, риторичних, але й граматичних ознак. Мета нашого дослідження 1) проаналізувати деякі типові лінгвістичні особливості заголовків, 2) виокремити проблеми, з якими зустрічаються перекладачі при перекладі заголовків новин. Тож завдання нашої розвідки: 1) розглянути лексичні та синтаксичні особливості заголовків електронних новин; 2) виокремити труднощі перекладу заголовків.

The relevance of our research is that chrematonyms are an integral part of the fiction text. They not only perform the nominative function, but also carry an additional semantic meaning. The research of semantic peculiarities of chrematonyms in fictional literature will help to analyze the sources of its formation.

The nature of the bilingual mental lexicon and how the L1 and the L2 interact in language production and processing has been the focus of decades of research from linguistic, psycholinguistic and neurolinguistic perspectives. In spite of this significant body of evidence, the degree to which the L1 influences L2 production and processing remains an area of debate, especially with respect to formal features such as grammatical gender. While it is clear that non-nativelike production and processing of L2 grammatical gender persist even in highly-proficient adult bilinguals, the underlying representation of the L1 and L2 gender features and how this representation affects the use of gender in the L2 is currently unclear. Furthermore, there is no evidence at present regarding the nature of the L1-L2 grammatical gender system when the L1 and the L2 have asymmetric gender systems (in other words, differ in number of gender values), as is the case with German, which bears three gender values (masculine, feminine and neuter), and Spanish and French, which each display two gender values (masculine and feminine). This dissertation investigates the representation of and interactions between the L1 and the L2 at the level of the formal gender feature, with a particular focus on language pairings with asymmetric gender systems. Through complementary data from L2 production and processing, I examine the representation of the asymmetric grammatical gender systems in the mental lexicon of L1 Spanish-L2 German and L1 French-L2 German bilinguals and the consequences this asymmetry between the L1 and L2 gender systems has on gender use strategies in the L2. From the perspective of bilingual lexical access, this research contributes new evidence to inform existing psycholinguistic theories of L1-L2 gender interactions and also proposes the Asymmetric gender representation hypothesis, a new model to account for the unique integrated nature of the gender system in bilinguals with L1-L2 asymmetric gender systems. From a language acquisition perspective, the present study provides new data on L2 gender use strategies with asymmetric gender systems, formulating the L1 transfer continuum, which extends existing proposals to include the degree of (a)symmetry between the L1 and the L2. This research also connects theoretical proposals regarding gender agreement in functional-lexical code-switches (specifically, switches within the Determiner Phrase such as dieGER-F mesaSPA-F or elSPA-M TischGER-M) to bilinguals’ preferences in code-switching between two languages that display formal gender. Taken together, all of these complementary perspectives addressed in this dissertation offer a well-rounded perspective of grammatical gender in asymmetric gender systems specifically, and contribute novel evidence regarding the interactions between the L1 and the L2 in the bilingual mental lexicon in general.

Об’єктом дослідження є англомовні друковані рекламні тексти. Предметом дослідження виступають їх лексико-семантичні особливості. Матеріалом дослідження є англомовні рекламні тексти провідних зарубіжних компаній. Мета роботи полягає у вивченні лексико-семантичних особливостей друкованих рекламних текстів.

Folklore plays an important role in studying social development, socioeconomics, customs and habits. Because language is a special level of vocabulary, studying folklore and language of fairy tale is crucial for understanding social development. Compared with other types of folk literature and art, fairy tales are the most popular and most extensive genre, which contains rich ideas accumulated by people over the centuries. Therefore, folklore and fairy tales are the basis for studying national psychology and personal characteristics.<br>В изучении развития общества, его экономики и обычаев, важное место занимает изучение фольклора, язык которого представляет собой особый пласт лексики, поэтому изучение языка фольклора и сказок. Имеет большое значение в понимании развития общества. По сравнению с другими видами фольклора, волшебная сказка является самым популярным и распространенным жанром, в котором содержатся богатые идеи, мысли, которые народ накопил в течение веков. Поэтому фольклор и сказки являются основой изучения национальной психологии и личностных характеристик.

Магістерська дисертація складається зі вступу, трьох розділів, висновків до кожного з них, загальних висновків та списку використаної літератури, який налічує 63джерела. Загальний обсяг роботи становить 107 сторінок. Предметом вивчення багатьох напрямків лінгвістики сьогодні виступає текст як структурно-смислове ціле. Він є основним засобом комунікації. Вивчення текстової зв'язності набуло актуальності у лінгвістиці тексту наприкінці 60-х років XXст. Зв'язність тексту здійснюється за допомогою низки структурних та лексико-семантичних засобів, які є у кожному тексті у найрізноманітніших комбінаціях, серед яких головним є повтор. Актуальність теми роботи зумовлена тим, що за останні десятиліття у науці накопичено досить ємний матеріал у галузі дослідження повтору, який потребує осмислення та певного узагальнення. Необхідно зазначити, що наразі недостатньо уваги приділено дослідженню особливостей функціонування лексичних повторів у художньому дискурсі. Незважаючи на те, що було написано велику кількість праць стосовно лексичних повторів, слід відзначити, що недостатньо проаналізовано питання про функціональну специфіку та розуміння повтору як стилістичного прийому або як елементу стилю письменника в межах художньої прози. Об’єктом дослідження є лексичні повтори у сучасній німецькій мові, а предметом — структурно-семантичні особливості їх функціонування у сучасній німецькій мові на основі художнього дискурсу німецькомовних письменників. Мета роботи полягає у дослідженні структурно-семантичних особливостей дискурсивного функціонування лексичних повторів у сучасній німецькій мові. Досягнення поставленої мети передбачає вирішення таких завдань: 1) вивчити поняття „повтор“; 2) розглянути типологічні різновиди повтору в сучасній німецькій мові; 3) описати лексичний повтор та його особливості; 4) схарактеризувати художній дискурс; 5) дослідити функціонування лексичних повторів у романі Даніеля Кельманна „Die Vermessung der Welt“; 6) визначити лексичні повтори у романі Даніеля Ґлаттауера „Gutgegen Nordwind“; 7) проаналізувати структурні особливості лексичних повторів у романі Акіфа Пірінчі „Der eine ist stumm, der andere ein Blinder“; 8) вивчити семантичні особливості лексичних повторів у романі Даніеля Кельманна; 9) розглянути лексичні повтори у романі Даніеля Ґлаттауера „Gutgegen Nordwind“; 10) здійснити дослідження семантики лексичних повторів у романі Акіфа Пірінчі „Der eine ist stumm, der andere ein Blinder“. Наукова новизна даної роботи полягає в тому, що вперше було проаналізовано особливості функціонування лексичного повтору на основі порівняння романів трьох авторів, романів двох німецьких письменників Даніеля Кельманна „Die Vermessung der Welt“ та Акіфа Пірінчі „Der eine ist stumm, der andere ein Blinder“, а також роману австрійського письменника Даніеля Ґлаттауера „Gutgegen Nordwind“. Практичне значення дослідження полягає у можливості використання матеріалів дослідження студентами вищих навчальних закладів, лінгвістичних відділень під час підготовки до семінарів, конференцій. Аналіз структурно-семантичних особливостей функціонування лексичних повторів у художньому дискурсі (на матеріалі романів сучасних німецькомовних авторів) може стати корисним при подальшому дослідженні лексичних повторів у прозових текстах та інших дискурсах. Матеріалом дослідження є 156 уривків текстів, у яких показано використання лексичних повторів, остаточна кількість яких сягає 200 одиниць, дібраних шляхом суцільної вибірки з 3 романів німецькомовних авторів, зокрема Даніеля Кельманна „Die Vermessung der Welt“, Даніеля Ґлаттауера „Gutgegen Nordwind“ та Акіфа Пірінчі „Der eine ist stumm, der andere ein Blinder“. Методи дослідження. Методика дослідження зумовлена метою та завданнями праці і ґрунтується на засадах комплексного застосування загальнонаукових методів: аналізу та синтезу – для відбору релевантного теоретичного матеріалу; індуктивно-дедуктивного методу – для спостереження, опису та класифікації лексичних повторів, кількісний метод – з метою встановлення співвідношень; методу класифікації та систематикації, методу зіставного аналізу і методу узагальнення науково-літературних джерел. Апробація результатів дослідження і публікації. Результати роботи висвітлено у статті «Функції лексичних повторів у текстах сучасної німецькомовної прози», опублікованій у науковому віснику Міжнародного гуманітарного університету серії «Філологія» (Вип. 39/2019, том 2) [29], який включено до Переліку фахових видань України та до міжнародної наукометричної бази Index Copernicus International (Республіка Польща).<br>The master's dissertation consists of an introduction, three chapters, conclusions to each chapter, general conclusions, and a list of references, that includes 63 points. The paper amounts to 107 pages. The subject of study of many areas of linguistics today is the text as a structural and semantic whole. It is the primary means of communication. The study of text coherence became relevant in text linguistics in the late 1960s. The cohesiveness of the text is achieved through a number of structural and lexical-semantic means, which are in each text in a variety of combinations, among which the main is repetition. The topicality is due to the fact that in recent decades in the science has accumulated enough capacious material in the field of repetition research, which needs reflection and some generalization. It should be noted, however, that insufficient attention has been paid to the study of the peculiarities of the functioning of lexical repetitions in artistic discourse. Although a great deal of work has been written on lexical repetitions, it is worth noting that the question of functional specificity and understanding of repetition as a stylistic reception or as an element of a writer's style within fiction is insufficiently analyzed. The object of the study is lexical repetition in modern German. The subject is the features of the functioning of lexical repetitions in modern German based on the artistic discourse of German-speaking writers. The aim of the research is to study the structural and semantic features of the discursive functioning of lexical repetitions in modern German. Achieving this goal involves the solution of the following tasks: 1) become familiar with the concept of "repetition"; 2) to consider typological varieties of repetition in modern German; 3) describe the lexical repetition and its features; 4) characterize artistic discourse; 5) to investigate the functioning of lexical repetitions in Daniel Kellmann's novel “Die Vermessung der Welt”; 6) identify lexical repetitions in Daniel Guttauer's novel “Gut gegen Nordwind”; 7) analyze the structural features of lexical repetitions in Akif Pirinchi's novel “Der eine ist stumm, der andere ein Blinder”; 8) to work out the semantic features of lexical repetition in Daniel Kelmann's novel; 9) consider lexical repetitions in Daniel Guttauer's novel “Gut gegen Nordwind”; 10) to carry out research on the semantic aspect of the study of lexical repetitions in Akef Pirinchi's novel “Der eineist stumm, der andere ein Blinder”. The originality of the research is that it was first analyzed the features of the functioning of lexical repetition on the basis of a comparison of novels of three authors, novels by two German writers Daniel Kellmann "Die Vermessung der Welt" and Akif Pirinchi "Der eine ist stumm, der andere ein Blinder" as well as the novel by Austrian writer Daniel Guttauer “Gut gegen Nordwind”.The practical valueof the research lies in the possibility of using research materials by students of higher education institutions, linguistic departments in preparation for seminars, conferences, and the analysis of the structural-semantic features of the functioning of lexical repetitions in the artistic discourse of German-speaking authors may be a useful example in the subsequent study of lexical studies in lexical studies. and other discourses. The research material amounts to 156 passages that show the use of lexical repetitions, the final number of which reaches 200 units selected by continuous sampling from 3 German-language novels, including Daniel Kellmann's “Die Vermessung der Welt”, Daniel Guttauer's “Gut gegen Nordwind” and Aqifa Pirinchi's “Der eine ist stumm, der andere ein Blinder“.Research methods. The research methodology is predetermined by the purpose and tasks of the work and is based on the principles of complex application of common scientific methods: analysis and synthesis - for the selection of relevant theoretical material; inductive-deductive method - for observing, describing and classifying lexical repetitions, quantitative method - for establishing repetitiveness, correlation; methods of classification and systematization, method of comparative analysis and method of generalization of scientific and literary sources. Testing the results of the research and publications. The results of the dissertation research are presented in the article "Functions of lexical repetition in modern German prose" in the scientific journal of the International Humanities University (Issue 39/2019, vol. 2) [29], which is notated in such international scientific metric database as Index Copernicus International (Poland).<br>Объектом исследования являются лексические повторы в современном немецком языке, а предметом – структурно-семантические особенности их функционирования в современном немецком языке на основе художественного дискурса немецкоязычных писателей. Цель работы заключается в исследовании структурно-семантических особенностей дискурсивного функционирования лексических повторов в современном немецком языке. Достижение поставленной цели предполагает решение следующих задач: изучить понятие "повтор"; рассмотреть типологические разновидности повтора в современном немецком языке; описать лексический повтор и его особенности; охарактеризовать художественный дискурс; исследовать функционирование лексических повторов в романе Даниэля Кельманна "Die Vermessung der Welt"; определить лексические повторы в романе Даниэля Ґлаттауэра "Gut gegen Nordwind"; проанализировать структурные особенности лексических повторов в романе Акифа Пиринчи "Der eine ist stumm, der andere ein Blinder"; изучить семантические особенности лексических повторов в романе Даниэля Кельманна; рассмотреть лексические повторы в романе Даниэля Ґлаттауэра "Gut gegen Nordwind"; провести исследование семантики лексических повторов в романе Акифа Пиринчи "Der eine ist stumm, der andere ein Blinder".

Pendant la perception de la parole, les locuteurs sont biaisés par un grand nombre de facteurs. Par exemple, il existe des limitations cognitives comme la mémoire ou l’attention, mais aussi des limitations linguistiques comme leur langue maternelle. Cette thèse se concentre sur deux de ces facteurs : les biais de traitement pendant la reconnaissance des mots, et les biais d’apprentissage pendant le processus de transmission. Ces facteurs peuvent se combiner et, au cours du temps, influencer l’évolution des langues. Dans la première partie de cette thèse, nous nous concentrons sur le processus de la reconnaissance des mots. Des recherches antérieures ont établi l’importance des traits phonologiques (p. ex. le voisement ou le lieu d’articulation) pendant le traitement de la parole. Cependant, nous en savons peu sur leur poids relatif les uns par rapport aux autres, et comment cela peut influencer la capacité des locuteurs à reconnaître les mots. Nous avons testé des locuteurs français sur leur capacité à reconnaître des mots mal prononcés et avons trouvé que les traits de mode et de lieu sont plus importants que le trait de voisement. Nous avons ensuite considéré deux sources de cette asymétrie et avons trouvé que les locuteurs sont biaisés et par la perception acoustique ascendante (les contrastes de mode sont plus facile à percevoir à cause de leur distance acoustique importante) et par la connaissance lexicale descendante (le trait de lieu est plus exploité dans le lexique français que les autres traits). Nous suggérons que ces deux sources de biais se combinent pour influencer les locuteurs lors de la reconnaissance des mots. Dans la seconde partie de cette thèse, nous nous concentrons sur la question d’un biais d’apprentissage. Il a été suggéré que les apprenants peuvent être biaisés vers l’apprentissage de certains patrons phonologiques grâce à leurs connaissances phonétiques. Cela peut alors expliquer pourquoi certains patrons sont récurrents dans la typologie, tandis que d’autres restent rares ou non-attestés. Plus spécifiquement, nous avons exploré le rôle d’un biais d’apprentissage sur l’acquisition de la règle typologiquement commune de l’harmonie vocalique comparée à celle de la règle non-attestée (mais logiquement équivalente) de la disharmonie vocalique. Nous avons trouvé des preuves d’un biais d’apprentissage aussi bien en perception qu’en production. En utilisant un modèle d’apprentissage itéré simulé, nous avons ensuite montré comment un biais, même petit, favorisant l’un des patrons, peut influencer la typologie linguistique au cours du temps et donc expliquer (en partie) la prépondérance de systèmes harmoniques. De plus, nous avons exploré le rôle du sommeil sur la consolidation mnésique. Nous avons montré que seul le patron commun bénéficie d’une consolidation et que cela est un facteur supplémentaire pouvant contribuer à l’asymétrie typologique. Dans l’ensemble, cette thèse considère certaines des sources de biais possibles chez l’individu et discute de comment ces influences peuvent, au cours du temps, faire évoluer les systèmes linguistiques. Nous avons démontré la nature dynamique et complexe du traitement de la parole, à la fois en perception et dans l’apprentissage. De futurs travaux devront explorer plus en détail comment ces différentes sources de biais sont pondérées les unes relativement aux autres<br>During speech perception, listeners are biased by a great number of factors, including cognitive limitations such as memory and attention and linguistic limitations such as their native language. This thesis focuses on two of these factors: processing bias during word recognition, and learning bias during the transmission process. These factors are combinatorial and can, over time, affect the way languages evolve. In the first part of this thesis, we focus on the process of word recognition. Previous research has established the importance of phonological features (e.g., voicing or place of articulation) during speech processing, but little is known about their weight relative to one another, and how this influences listeners' ability to recognize words. We tested French participants on their ability to recognize mispronounced words and found that the manner and place features were more important than the voicing feature. We then explored two sources of this asymmetry and found that listeners were biased both by bottom-up acoustic perception (manner contrasts are easier to perceive because of their acoustic distance compared to the other features) and top-down lexical knowledge (the place feature is used more in the French lexicon than the other two features). We suggest that these two sources of bias coalesce during the word recognition process to influence listeners. In the second part of this thesis, we turn to the question of bias during the learning process. It has been suggested that language learners may be biased towards the learning of certain phonological patterns because of phonetic knowledge they have. This in turn can explain why certain patterns are recurrent in the typology while others remain rare or unattested. Specifically, we explored the role of learning bias on the acquisition of the typologically common rule of vowel harmony compared to the unattested (but logically equivalent) rule of vowel disharmony. We found that in both perception and production, there was evidence of a learning bias, and using a simulated iterated learning model, showed how even a small bias favoring one pattern over the other could influence the linguistic typology over time, thus explaining (in part) the prevalence of harmonic systems. We additionally explored the role of sleep on memory consolidation and showed evidence that the common pattern benefits from consolidation that the unattested pattern does not, a factor that may also contribute to the typological asymmetry. Overall, this thesis considers a few of the wide-ranging sources of bias in the individual and discusses how these influences can over time shape linguistic systems. We demonstrate the dynamic and complicated nature of speech processing (both in perception and learning) and open the door for future research to explore in finer detail just how these different sources of bias are weighted relative to one another

Large variations in writing styles and difficulties in segmenting cursive words are the main reasons for handwritten cursive words recognition for being such a challenging task. An Indian postal document reading system based on a segmentation-free context based stochastic model is presented. The originality of the work resides on a combination of high-level perceptual features with the low-level pixel information considered by the former model and a pruning strategy in the Viterbi decoding to reduce the recognition time. While the low-level information can be easily extracted from the analyzed form, the discriminative power of such information has some limits as describes the shape with less precision. For that reason, we have considered in the framework of an analytical approach, using an implicit segmentation, the implant of high-level information reduced to a lower level. This enrichment can be perceived as a weight at pixel level, assigning an importance to each analyzed pixel based on their perceptual properties. The challenge is to combine the different type of features considering a certain dependence between them. To reduce the decoding time in the Viterbi search, a cumulative threshold mechanism is proposed in a flat lexicon representation. Instead of using a trie representation where the common prefix parts are shared we propose a threshold mechanism in the flat lexicon where based just on a partial Viterbi analysis, we can prune a model and stop the further processing. The cumulative thresholds are based on matching scores calculated at each letter level, allowing a certain dynamic and elasticity to the model. As we are interested in a complete postal address recognition system, we have also focused our attention on digit recognition, proposing different neural and stochastic solutions. To increase the accuracy and robustness of the classifiers a combination scheme is also proposed. The results obtained on different datasets written on Latin and Bengali scripts have shown the interest of the method and the recognition module developed will be integrated in a generic system for the Indian postal automation.

Notre étude s'inscrit dans le cadre de la théorie des matrices et des étymons (TME), principalement élaborée par G. Bohas (1997, 2000), G. Bohas et M. Dat (2007) et G. Bohas et A. Saguer (2012). Ce nouvel outil propose une réorganisation du lexique des langues sémitiques non plus sur la base de phonèmes mais de traits phonétiques. Cette perspective mène à contester le caractère primitif de la notion de racine triconsonantique développée par les grammairiens arabes du Moyen-Âge. De plus, la TME permet de rendre compte d'un certain nombre de régularités observées dans le lexique, telles que les liens phonético-sémantiques existants entre certains radicaux, l'aspect mimétique de la structure du signe, la polysémie des racines trilitères, etc. Notre thèse traite dans ce cadre du vocabulaire de l'hébreu biblique et se présente en trois parties. Dans un premier temps est donnée une description complète du fonctionnement de la théorie, suit un développement du vocabulaire de sept champs notionnels contraints par un cadre phonétique stable puis nous proposons un dictionnaire présentant une réorganisation totale du lexique hébraïque ancien sur la base d'étymons bilitères.

Studies in Genre Analysis have been developed with the aim of exploring the typification of genres of the work environment and interpreting their funtions according to the discursive communities in which they operate. Some examples are studies carried out by Bazerman (2005) and Devitt (1991) on documents from professional contexts. In that regard, Bazerman (2005) discusses notions such as genre, genre sets, genre systems and activity systems. Language practices in professional contexts have also been the object of investigations among Brazilian research teams, for instance LAEL at PUC/SP and PUC/RJ, in efforts to understand how the participants of a communicative event interact and how their lexicogrammatical choices construe meaning. In that sense, the present study approaches the system of genres adopted in the selective process for hiring a salesperson, conducted by a cooperative company in the state of Rio Grande do Sul. This system of genres comprises several texts: Want Ad, Presentation Letter, Curriculum Vitae and Interview. The objective of this study is to investigate the system of genres for job application from the perspective of language as social practice with a focus on practices circumsbribed to the work sphere. This way, the struture of the texts within this system of genres is analyzed using the framework of Systemic-Functional Linguistics (Halliday, 1994; 2004 and Halliday & Hasan, 1989). The texts are analyzed in their organizational patterns and clauses are classified considering the type of processes, according to the System of Transitivity. Besides, the communicative purposes as welll as the characterization of the participants in the activities system are identified. Contrary to initial expectations, the interview was pointed out in the results as the most relevant genre in evaluating the applicants. Overall, the results offer an understanding of the interaction as characterized by specific aims in a particular professional sphere and can be valid for language educators, mainly those who work with Portuguese as mother tongue for instrumental purposes with a focus on professional genres.<br>Estudos em Análise de Gênero têm sido desenvolvidos a fim de explorar a tipificação de gêneros do contexto do trabalho e interpretar suas funções de acordo com as comunidades discursivas em que operam. Alguns exemplos são os estudos desenvolvidos por Bazerman (2005) e Devitt (1991) sobre documentos de contextos profissionais. Em vista disso, Bazerman (2005) discute noções sobre gênero, conjunto de gêneros, sistema de gêneros e sistema de atividades. A linguagem voltada a contextos profissionais também tem sido objeto de análise para grupos de pesquisa brasileiros, por exemplo, o do LAEL da PUC/SP e o da PUC/RJ, no sentido de compreender como os participantes de um evento comunicativo interagem e como suas escolhas léxico-gramaticais constroem significado. Assim, o tema do presente estudo diz respeito ao sistema de gêneros de um processo seletivo para contratação de vendedor autônomo, realizado em uma empresa cooperativa do interior do Rio Grande do Sul. Esse sistema de gêneros acomoda vários textos: Anúncio de Emprego, Carta de Apresentação, Curriculum Vitae e Entrevista Pessoal. O objetivo deste estudo é investigar o sistema de gêneros da seleção a emprego a partir da perspectiva de linguagem como prática social, com ênfase nas práticas relacionadas às esferas do trabalho. Desse modo, a estrutura dos exemplares desse sistema de gêneros é descrita a partir de propostas da Lingüística Sistêmico-Funcional (Halliday, 1994; 2004 e Halliday e Hasan, 1989). São identificados os padrões de organização dos textos e classificados os processos das orações, conforme o Sistema de Transitividade. E, são identificados os propósitos comunicativos e o perfil dos participantes desse sistema de atividades. Na avaliação dos candidatos, contrariando a expectativa inicial, os resultados destacam a relevância maior para a Entrevista Pessoal. Enfim, a discussão aponta à compreensão de como se dá a interação, caracterizada por propósitos específicos em uma situação profissional e pode ser válida para os profissionais da linguagem, principalmente, àqueles que trabalham com o ensino instrumental da língua materna, voltado aos gêneros do contexto empresarial.

碩士<br>國立臺北大學<br>資訊管理研究所<br>100<br>Personal computer computation power and connection capability dramatically increase such that personal computers becomes malicious attack major targets instead of traditional servers. By using worms and Trojan horses infecting victim personal computers, attackers establish their Botnets, which remote control victims performing malicious activities in order to make money and thus become major Internet threats. Encryption and domain fluxing become current major evading techniques for Botnet. In order to defend Botnet, DNS black list is one of the most effective defensive strategies to hidden Botnet connections. Therefore, effectively identifying malicious domain names is a critical issue for Botnet detection. This paper presents lexical analysis approach to find domain name different patterns, and adopts decision tree models to train the best combination of malicious domain name lexical features. We propose five major features: domain length, syllable count, vowel count, vowel ratio, character redundant. Experiment results show vowel ratio and vowel count can effectively differentiate black and white list samples while cross comparing them. While calculating false positive and negative rates, the combination of vowel ratio and vowel count provides the best results, 0.01% false positive rate and 4.3 % false negative rate, which show our approach can effectively identify malicious domain names.

碩士<br>國立中央大學<br>資訊工程學系在職專班<br>107<br>Sequence labeling model has been widely used in Natural Language Processing (NLP). Ex: Named Entity recognition (NER), Part-Of-Speech tagging (POS) and Word Segmentation. Named Entity Recognition (NER) is one of the important tasks of Natural Language Processing because it can extract unnamed articles and extract them into pre-defined categories, such as person name, place name, organization, etc. Most of the research in Named Entity Recognition (NER) focused on English data. In English, spaces are usually used for dividing words, and each word has its own meaning. While in Chinese, each characters contains different information, different location of the vocabulary, may represent different meanings, so Chinese is without explicit word delimiters. However, the traditional machine learning of Chinese Named Entity Recognition (CNER), most of them use statistical methods and take the Conditional Random Field (CRF) to complete the sequence labeling task. Therefore, it only can capture local features. It is a challenging and forward-looking task to capturing long-range context information in Chinese dataset, determine the correct semantic meaning of the current word, and correctly identify the named entity. In order to overcome the challenges, this study used the deep learning Condition Random Fields to execute Chinese Named Entity Recognition task. Firstly, training a word vector model to convert characters to numeric data. And used convolutional layer, bidirectional GRU layer, and the memory layer that integrates external memory contains long-range context information. Making the task different from usual, only can capture local information, but can obtain rich message of article. Also by feature extraction generate some lexical features[1]. And use a automatically trained variable of deep learning model to automatically adjust the weight of word embedding and lexical features. In addition of long-range article information, the model also can fully obtain the hidden information of article. The data set used in this research includes PerNews which is online articles collected using custom crawler as training data and online news articles as test data, and SIGHAN Bakeoff-3. According to the results, the model proposed in this research achieve 91.67% tagging accuracy in the online social media data. The result is significantly higher than the model that doesn’t add memory layer by 2.9%. And then the word embedding and lexical features are added, compared with the basic memory model increase 6.04%. The model proposed in this study also achieve the highest F1-score 92.45% at location name entity recognition performance and 90.95% overall recall rate in SIGHAN-MSRA dataset.

碩士<br>國立臺灣大學<br>華語教學碩士學位學程<br>105<br>Newspaper is an important print media using succinct style to convey a large amount of information; it has unique register features in terms of the structure, syntax and lexicons. As one of the useful authentic materials in language teaching, newspaper can expose students not only to the real language used by native speakers but also to the current issues in the local society and culture. Although newspaper is commonly regarded as an important source of reading materials in language courses, students often find it difficult to read newspaper articles as newspaper covers a wide range of topics and vocabulary. This study aims to investigate the vocabulary of the newspaper register by creating the specialized Chinese newspaper word list and discussing vocabulary features based on a collection of 900 newspaper texts of total 619,494 characters. Researchers has put forward the concept of “word list” to set vocabulary goals for different leaning purposes. Furthermore, word list can provide learners with a reference to first focus on essential words and promote students’ learning efficiency. The current study collected 900 newspaper texts published from January to March in 2016 in three major newspapers in Taiwan: China Times, Liberty Times and United Daily News. The newspaper texts collected and examined come from five news sections: World news, Culture and Education, Business, Entertainment and Lifestyle. This study first came up with words of high frequency in the newspaper and then excluded the basic words to create the specialized Chinese Newspaper word list of 530 words. The lexical features of the vocabulary in the specialized word list were discussed. Besides, the high frequency words in each of the five news sections were also listed and discussed. Based on the results, the design for newspaper reading courses and teaching suggestions are proposed, which hopefully could provide some guidelines for learners and material designers in selecting texts.

碩士<br>國立臺北大學<br>資訊工程學系<br>106<br>The latest threat reports show a notable increase in detected botnets compared to previous years. In fact, the number of IoT botnet C&C controllers alone more than doubled in 2017. Botnet C&C controllers are used by cybercriminals to launch attacks using botnet enslaved devices. As showed success of Mirai botnet, a lot of companies use poorly secured IoT devices, which gives an opportunity for using IoT devices as botnet zombies. To avoid detection, botnets use domain generation algorithms (DGA) to connect to C&C servers via large number of domain names. This work proposes a low-cost strategy to detect domain names generated by DGA. Statistically lexical features of domain names generated algorithmically differ from those generated by humans. Thus, algorithmically generated domain names can be used to detect botnet or malware activity in the network. To justify the choice of lexical features we gathered domain names statistics of 32 botnets that appeared in last 8 years. Lexical features were chosen based on gathered statistics and new lexical features were suggested. Chosen lexical features were used to generate a decision tree by means of C4.5 algorithm. Experimental results show, that suggested new lexical features improve detection accuracy. 93.7% detection accuracy was achieved. Detection algorithm based on the generated decision tree can be used for fast real-time detection of botnet domain names.

碩士<br>國立高雄師範大學<br>華語文教學研究所<br>102<br>The communication between Chinese and English characters is over 200 years. Its nonconformity in concept of Chinese characters and English words is the background of this research. Therefore it is essential to scrutinize Chinese characters and English words again. The purpose of this study is to research the internal features of lexicon in Chinese characters that keep the same with English word concept. This is a qualitative research, which begins from August of the 99th year of the Republic of China to February of 103. Data are from two English texts, 124 Chinese written materials, and 100 classified Chinese characters according to its onomatopoeia. Data analysis includes the following three steps: first, analyzing the present word-building situation of Chinese and English; second, comparing the Chinese characters with English words in order to identify ' the word ' unit of Chinese; third, synthesizing the features of lexicon showed in Chinese characters. Two research findings are discussed. First, the word concept of English includes that its presentation of independent meaning is the independent sound and shape, word includes the independent morpheme, characteristics are a bunch of language forms, and its morpheme form changes depending upon building word. Second, the internal feature of lexicon in Chinese characters keeps the same with English word concept, which includes its presentation of independent meaning is the independent sound and shape, word includes the independent morpheme, characteristics is a piece of language form, and the morpheme form changes depending upon building word. In conclusion, Chinese characters are the word-characters. Comparing to English, its presentations of the internal feature of lexicon in Chinese characters are: First, the word symbol of Chinese language is a single Chinese character; Second, the Chinese character phenomenon of word is one word sound which is appropriate to bind with many morphemic sounds. The researcher then suggests that future studies in word-building of Chinese should begin its inquiry from Chinese-character-word.

博士<br>國立臺灣師範大學<br>華語文教學系<br>101<br>The study aims to explore the salient linguistic features of Chinese lexical items and syntactic structures across the CEFR levels. The research based on corpus linguistics and contrastive interlanguage analysis method, including comparing learner corpus (i.e. L2 corpus) and native-speaker corpus (i.e. L1 corpus), as well as different native language background L2 sub-corpora. The L2 corpus which consists of more than 1.14 million Chinese words from novice proficiency to advanced learners’ texts is mainly from the computer-based writing Test of Chinese as a Foreign Language (TOCFL). L1 corpus is from the Academia Sinica balanced corpus. In lexical part, we try to seek possible criterial features (Hawkins &; Buttery, 2009) from the aspects of vocabulary size, high frequency information, error rate of word use and keyword-keyness analysis. Some salient features are discovered. For examples, English L2 shows the unusual high frequency on pronouns and unusual low frequency on sentential final particles in Chinese writing. And Japanese as well as Korean L2s tend to overuse the post form ‘de hua’ when expressing the If sentence, and overuse ‘suoyi’ when expressing the cause-result relation. The dissertation also provides possible explanations for the findings from the aspects of L2s’ native language typology, linguistic structure, syntactic category and culture. In structure part, we observe and discuss the syntactic features from the aspects of nominal phrase structure, frequent patterns and word order expression by different proficiency learners. We find the A2 level learners have acquired the canonical Chinese word order, for example, SVO. The specific structures like Ba sentence, Bei sentence, and Lian sentence are proved as the criterial features to discriminate the different proficiency levels. And the features of frequency of using relative clause and the expansion ability of nominal phrase are also discriminative for Chinese learners.

Ho, Pui Kwan.<br>"November 2010."<br>Thesis (M.Phil.)--Chinese University of Hong Kong, 2011.<br>Includes bibliographical references (leaves 68-75).<br>Abstracts in English and Chinese.<br>Acknowledgements --- p.i<br>Abstract --- p.ii<br>List of figures --- p.vi<br>List of tables --- p.viii<br>Abbreviations --- p.ix<br>Chapter Chapter 1. --- Introduction<br>Chapter 1.1. --- Overview --- p.1<br>Chapter 1.2. --- Thesis organization --- p.6<br>Chapter Chapter 2. --- Background<br>Chapter 2.1. --- Brain structure and its functional organization --- p.8<br>Chapter 2.2. --- Introduction to lateralization investigation techniques --- p.13<br>Chapter 2.2.1. --- Lesion studies and Wada test --- p.13<br>Chapter 2.2.2. --- Dichotic listening --- p.14<br>Chapter 2.2.3. --- Functional Magnetic Resonance Imaging (fMRI) --- p.15<br>Chapter 2.2.4. --- Electroencephalography(EEG) & event-related potential(ERP) --- p.16<br>Chapter 2.2.5. --- Other brain imaging techniques --- p.20<br>Chapter 2.2.6. --- Comparing different brain-imaging techniques in this study --- p.20<br>Chapter 2.3. --- Literature review and the current work --- p.21<br>Chapter 2.3.1 --- "The functional and acoustic ""dichotomy""" --- p.22<br>Chapter 2.3.2. --- Studies on lateralization of tones without linguistic context --- p.25<br>Chapter 2.3.3. --- Studies on lexical tone lateralization --- p.27<br>Chapter 2.3.4. --- The current work ´ؤ aim and hypothesis --- p.34<br>Chapter Chapter 3. --- Methodology<br>Chapter 3.1. --- Stimuli --- p.36<br>Chapter 3.2. --- Subject --- p.41<br>Chapter 3.3. --- Task design --- p.41<br>Chapter 3.4. --- ERP recording and data processing --- p.44<br>Chapter Chapter 4. --- Results and discussions<br>Chapter 4.1. --- Behavioural data analysis --- p.46<br>Chapter 4.2. --- ERP data analysis --- p.48<br>Chapter 4.3. --- Discussions --- p.52<br>Chapter Chapter 5. --- Further discussions<br>Chapter 5.1. --- Is lateralization of lexical tone language-specific? --- p.57<br>Chapter 5.2. --- Toward a model of speech lateralization --- p.59<br>Chapter 5.3. --- An implication on tone feature systems --- p.60<br>Chapter Chapter 6. --- Conclusion<br>Chapter 6.1. --- Conclusion --- p.65<br>Chapter 6.2. --- Future work --- p.67<br>References --- p.68<br>Appendix<br>Chapter A.1. --- Edinburgh handedness test --- p.76<br>Chapter A.2. --- Handedness scores --- p.77

碩士<br>國立臺北護理健康大學<br>聽語障礙科學研究所<br>101<br>Purpose: To investigate the production of lexical tones and intonation in adolescent native speakers of Taiwan Mandarin with High-Functioning Autism (HFA) or Asperger’s Syndrome (AS) and compare to the typical adolescents. The main goals of the current study were (1) to describe the production of the four lexical tones in syllables and in sentences through acoustic parameters; and (2) to plot the pitch contours of lexical tones in different utterance types, i.e. word in isolation, statement and question. Method: Ten adolescents with HFA or AS and ten typical adolescents participated in 3 tasks that examined the productions of four lexical tones in syllable, statements, and questions. In the tasks involving statements and questions, the tone variations were introduced either in the 3rd or in the 5th syllable. Productions were analyzed with acoustic parameters – mean pitch, intensity, duration, pitch range, pitcht10-pitcht1 and pitch slope. The time-normalised pitch of each syllable was used to plot pitch contour for different lexical tones, utterance types (state-ment vs. questions) and syllable position in utterance (3rd vs. 5th syllable). Results: No group differences were found in any of the investigated acoustic parameters. When considering pitch contours, although both groups had similarities in the shape, the pitch contour lines for different lexical tones in different utterance types and in different syllable position in utterance were more ‘packed together’ for adolescents with HFA or AS. Conclusion: Adolescents with HFA or AS were as capable as their typical peers in producing lexical tones in different contexts, i.e., either in different utterance types or in different sylla-ble positions in utterance.

碩士<br>國立交通大學<br>統計學研究所<br>106<br>The thesis utilize the financial lexicon and text information in Form 10-k to predict the post-event volatility after the release date of Form 10-k. We also propose some methods in feature selection and transformation to imporove the prediction accuracy. Furthurmore, we examine the prediction capability of the selected text features by introducing GARCH.

Удовенко А. Д. Англомовний бізнесовий дискурс та особливості його перекладу (на матеріалі бізнесових презентацій) : кваліфікаційна робота магістра спеціальності 035 «Філологія» / наук. керівник І. О. Андрєєва. Запоріжжя : ЗНУ, 2020. 90 с.<br>UA : Дипломна робота − 90 стор., 114 джерел. Об’єкт дослідження: англомовний дискурс бізнесових презентацій. Мета роботи: визначення особливостей перекладу англомовних бізнесових презентацій українською і російською мовами. Теоретико-методологічні засади: ключові положення дискурсології та перекладознавства розроблені В. Демецькою, В. Карасиком, О. Морозовою, принципи теорії перекладу, розроблені В. Карабаном, І. Корунцем, Т. Кияком, К. Райс та ін.). Отримані результати: Найменшою одиницею презентаційного дискурсу є презентема – найдрібніша інформаційна одиниця впливу, що є складним лінгвосеміотичним (знаковим) комплексом, складається з когнітивно-засвоєних суб’єктом концептів та образів навколишнього світу і переданий іншому суб'єкту в ході комунікації з цим суб'єктом з метою впливу на нього. Ключовими концептами бізнесових презентацій є Failure, Success, Cooperation і Competition, які накладаються утворюючи концептуальну систему бізнесової презентації. Периферійні концепти, що реалізуються в межах однієї концептосистеми це – Leadership, Customer, Rival, Threat, Persuasion, Profit, Aim. Під час перекладу бізнесових презентацій, перекладач вирішує велику кількість перекладацьких проблем, загал яких можна розподілити на три великих групи: граматичні особливості, лексичні особливості, стилістичні особливості презентаційного дискурсу.<br>EN : Thesis - 90 pages, 114 sources. Object of research: English discourse of business presentations. Purpose: to determine the features of English translation business presentations in Ukrainian and Russian. Theoretical and methodological principles: key provisions of discourse and translation studies developed by V. Demetskaya, V. Karasyk, O. Morozova, principles of translation theory, developed by V. Karaban, I. Korunets, T. Kiyak, K. Rice, etc.). Results obtained: The smallest unit of presentation discourse is a present - the smallest information unit of influence that exists complex linguosemiotic (symbolic) complex, consists of cognitively mastered by the subject concepts and images of the world around and transferred to another entity in the course of communication with that entity for the purpose impact on him. The key concepts of business presentations are Failure, Success, Cooperation and Competition, which overlap forming conceptual system of business presentation. Peripheral concepts that implemented within one concept system is - Leadership, Customer, Rival, Threat, Persuasion, Profit, Aim. When translating business presentations, the translator solves a large number of translation problems, which can be divided into three major groups: grammatical features, lexical features, stylistic features of presentation discourse.