Academic literature on the topic 'Structure dissimilarity'

Clustering is a very useful technique to categorise Web users with common browsing activities, access patterns and navigational behaviour. Web user clustering is used to build Web visitor profiles that make the core of a personalised information recommender system. These systems are used to comprehend Web users surfing activities by offering tailored content to Web users with similar interests. The principle objective of Web user sessions clustering is to maximise the intra-group while minimising the inter-group similarity. Efficient clustering of Web users’ sessions not only depend on the clustering algorithm’s nature but also depend on how well user concerns are captured and accommodated by the dissimilarity measure that are used. Determining the right dissimilarity measure to capture the access behaviour of the Web user is very significant for substantial clustering. In this paper, an intuitive dissimilarity measure is presented to estimate a Web user’s concern from augmented Web user sessions. The proposed usage dissimilarity measure between two Web user sessions is based on the accessing page relevance, the syntactic structure of page URL and hierarchical structure of the website. This proposed intuitive dissimilarity measure was used with K-Medoids Clustering algorithm for experimentation and results were compared with other independent dissimilarity measures. The worth of the generated clusters were evaluated by two unsupervised cluster validity indexes. The experimental results show that intuitive augmented session dissimilarity measure is more realistic and superior as compared to the other independent dissimilarity measures regarding cluster validity indexes.

The development of analysis methods for categorical data begun in 90's decade, and it has been booming in the last years. On the other hand, the performance of many of these methods depends on the used metric. Therefore, determining a dissimilarity measure for categorical data is one of the most attractive and recent challenges in data mining problems. However, several similarity/dissimilarity measures proposed in the literature have drawbacks due to high computational cost, or poor performance. For this reason, we propose a new distance metric for categorical data. We call it: Weighted pairing (W-P) based on feature space-structure, where the weights are understood like a degree of contribution of an attribute to the compact cluster structure. The performance of W-P metric was evaluated in the unsupervised learning framework in terms of cluster quality index. We test the W-P in six real categorical datasets downloaded from the public UCI repository, and we make a comparison with the distance metric (DM3) method and hamming metric (H-SBI). Results show that our proposal outperforms DM3 and H-SBI in different experimental configurations. Also, the W-P achieves highest rand index values and a better clustering discriminant than the other methods.

This article contains asymmetric dissimilarity data which is observed in various situations. In asymmetric dissimilarity data, dissimilarity from subject i to j and from subject j to i are not the same necessarily. Asymmetric multidimensional scaling (AMDS) is a visualization method for describing the asymmetric relations between subjects, given asymmetric dissimilarity data for subjects. It is sure that AMDS is a useful tool for interpreting the asymmetric relation, however, existing AMDS cannot be considered for the external information, even if the external information of the same subjects for the asymmetric dissimilarity data is given. If the estimated coordinates can be interpreted from the loading matrix for the external information like principal component analysis (PCA), the AMDS become more useful. This is because we can interpret the relation between the estimated asymmetries and the factors of the external information on the low dimensions. In this article, we proposed new AMDS with external information. In addition to that, the proposed method can consider the path structure for variables like SEM.

Topographic maps such as the self-organizing map (SOM) or neural gas (NG) constitute powerful data mining techniques that allow simultaneously clustering data and inferring their topological structure, such that additional features, for example, browsing, become available. Both methods have been introduced for vectorial data sets; they require a classical feature encoding of information. Often data are available in the form of pairwise distances only, such as arise from a kernel matrix, a graph, or some general dissimilarity measure. In such cases, NG and SOM cannot be applied directly. In this article, we introduce relational topographic maps as an extension of relational clustering algorithms, which offer prototype-based representations of dissimilarity data, to incorporate neighborhood structure. These methods are equivalent to the standard (vectorial) techniques if a Euclidean embedding exists, while preventing the need to explicitly compute such an embedding. Extending these techniques for the general case of non-Euclidean dissimilarities makes possible an interpretation of relational clustering as clustering in pseudo-Euclidean space. We compare the methods to well-known clustering methods for proximity data based on deterministic annealing and discuss how far convergence can be guaranteed in the general case. Relational clustering is quadratic in the number of data points, which makes the algorithms infeasible for huge data sets. We propose an approximate patch version of relational clustering that runs in linear time. The effectiveness of the methods is demonstrated in a number of examples.

Since computing semantic similarity tends to simulate the thinking process of humans, semantic dissimilarity must play a part in this process. In this paper, we present a new approach for semantic similarity measuring by taking consideration of dissimilarity into the process of computation. Specifically, the proposed measures explore the potential antonymy in the hierarchical structure of WordNet to represent the dissimilarity between concepts and then combine the dissimilarity with the results of existing methods to achieve semantic similarity results. The relation between parameters and the correlation value is discussed in detail. The proposed model is then applied to different text granularity levels to validate the correctness on similarity measurement. Experimental results show that the proposed approach not only achieves high correlation value against human ratings but also has effective improvement to existing path-distance based methods on the word similarity level, in the meanwhile effectively correct existing sentence similarity method in some cases in Microsoft Research Paraphrase Corpus and SemEval-2014 date set.

Previous literature has found that individuals who are highly identified with their own ingroup and who perceive an outgroup as being more cohesive (viz., high in entitativity) will exhibit greater levels of vicarious retribution. Further studies have shown that an individual who perceives the outgroup as having dissimilar values to their own will engage in higher levels of direct aggression. The current study was the first to investigate value dissimilarity as a moderator in the relationship between both ingroup identification and outgroup entitativity on subsequent vicarious retribution. This study used a 3 (value dissimilarity: high, low, neutral/no information) x 2 (outgroup entitativity: high, low) between subjects design. Results indicated that among participants who had a strong reaction to the provocation, low value dissimilarity significantly lowered aggression thus serving to buffer the effect of provocation on subsequent vicarious retribution. Implications for reducing intergroup violence and vicarious retribution are discussed.

Of the 20 standard amino acids, cysteines are the only amino acids that have a reactive sulphur atom, thus enabling two cysteines to form strong covalent bonds known as disulfide bonds. Even though almost all proteins have cysteines, not all of them have disulfide bonds. Disulfide bonds provide structural stability to proteins and hence are an important constraint in determining the structure of a protein. As a result, disulfide bonds are used to study various protein properties, one of them being protein folding. Protein structure prediction is the problem of predicting the three-dimensional structure of a protein from its one-dimensional amino acid sequence. Ab initio methods are a group of methods that attempt to solve this problem from first principles, using only basic physico-chemical properties of proteins. These methods use structure libraries of short amino acid fragments in the process of predicting the structure of a protein. The protein structures from which these structure libraries are created are not classified in any other way apart from being non-redundant. In this thesis, we investigate the structural dissimilarities of short amino acid fragments when occurring in proteins with disulfide bonds and when occurring in those proteins without disulfide bonds. We are interested in this because, as mentioned earlier, the protein structures from which the structure libraries of ab initio methods are created, are not classified in any form. This means that any significant structural difference in amino acids and short fragments when occurring in proteins with and without disulfide bonds would remain unnoticed as these structure libraries have both fragments from proteins with disulfide bonds and without disulfide bonds together. Our investigation of structural dissimilarities of amino acids and short fragments is done in four phases. In phase one, by statistically analysing the phi and psi backbone dihedral angle distributions we show that these fragments have significantly different structures in terms of dihedral angles when occurring in proteins with and without disulfide bonds. In phase two, using directional statistics we investigate how structurally different are the 20 different amino acids and the short fragments when occurring in proteins with and without disulfide bonds. In phase three of our work, we investigate the differences in secondary structure preference of the 20 amino acids in proteins with and without disulfide bonds. In phase four, we further investigate and show that there are significant differences within the same secondary structure region of amino acids when they occur in proteins with and without disulfide bonds. Finally, we present the design and implementation details of a dihedral angle and secondary structure database of short amino acid fragments (DASSD) that is publicly available. Thus, in this thesis we show previously unknown significant structure differences in terms of backbone dihedral angles and secondary structures in amino acids and short fragments when they occur in proteins with and without disulfide bonds.

Les modèles d'occupation de niche au sein des communautés locales, la variabilité spatiale de la biodiversité le long des gradients environnementaux du stress et des perturbations, et les processus de succession végétale sont plusieurs sujets fondamentaux en écologie. Récemment, l'approche basée sur les traits est apparue comme un moyen prometteur de comprendre les processus structurant les communautés végétales et cette approche a même été proposée comme méthode pour reconstruire l'écologie communautaire en fonction des traits fonctionnels. Par conséquent, lier ces thèmes fondamentaux en utilisant une lentille fonctionnelle devrait nous donner un aperçu de certaines questions fondamentales en écologie et sera l'objectif principal de ma thèse. En général, mon projet de doctorat vise à étudier les structures de l'occupation de l'espace fonctionnel dans les gradients spatio-temporels. Plus précisément, l'objectif du chapitre 2 est (i) d'étudier les modèles d'occupation de la niche fonctionnelle en calculant trois métriques clés de niche (le volume total de niche fonctionnelle , le chevauchement des niches fonctionnelles et le volume de niche fonctionnel moyen) des communautés pauvres en espèces aux communautés riches en espèces et (ii) de déterminer le principal facteur de la structure observée de l'occupation de la niche fonctionnelle dans les communautés végétales du monde entier. Dans le chapitre 3, je vise à prédire et à expliquer la variation de la richesse en espèces selon les gradients de stress et de perturbation, en reliant le modèle d'équilibre dynamique et l'occupation de la niche fonctionnelle en fonction du cadre développé au chapitre 2. L'objectif du chapitre 4 est de tester expérimentalement l’application d'une méthode d'ordination CSR évaluée globalement en fonction de trois traits de feuilles (surface foliaire, teneur en matière sèche des feuilles et surface foliaire spécifique) dans les études locales. Enfin, l'objectif du chapitre 5 est de tester expérimentalement les hypothèses qui concilient les points de vue déterministes et historiquement contingents de la succession végétale, en étudiant la variation des divergences taxonomiques et fonctionnelles entre les communautés selon des gradients de stress et de perturbation. L'étude globale (chapitre 2) est basée sur une collection de 21 jeux de données, couvrant les biomes tropicaux et tempérés, et se compose de 313 communautés végétales représentant différentes formes de croissance. Les études locales (chapitre 3, 4 et 5) sont basées sur le même système expérimental constitué de 24 mésocosmes présentant différents niveaux de stress et de perturbation. L'expérience a commencé en 2009 avec le même mélange de graines de 30 espèces herbacées semées sur les 24 mésocosmes et s'est terminée en 2016. Nous avons permis la colonisation naturelle de graines de la banque commune de graines de sol et de l'environnement pendant la succession de sept ans. Dix traits ont été mesurés sur cinq individus (échantillonnés directement à partir des mésocosmes) par espèce par mésocosme en 2014 (chapitre 3 et 4). Un autre ensemble de traits (16 traits, y compris certains traits qui ne pouvaient pas être mesurés directement dans les mésocosmes), ont été mesurés au niveau de l'espèce (valeurs moyennes des traits) pour les 34 espèces les plus abondantes (certaines espèces disparues dans les mésocosmes) au cours des sept Ans, en les regroupant séparément pour une saison de croissance. Au chapitre 2, nous avons constaté que les communautés étaient plus diverses en termes fonctionnels (une augmentation du volume fonctionnel total) dans les communautés riches en espèces et que les espèces se chevauchaient davantage au sein de la communauté (augmentation du chevauchement fonctionnel), mais ne divisaient pas plus finement l'espace fonctionnel (aucune réduction du volume fonctionnel moyen). En outre, le filtrage de l'habitat est un processus répandu qui conduit à la caractérisation de l'occupation de niche fonctionnelle dans les communautés végétales. Dans le chapitre 3, nous avons trouvé un modèle similaire d'occupation de niche fonctionnelle sur un système expérimental avec une taille spatiale communautaire constante et un effort d'échantillonnage des traits, qui, avec le chapitre 2, nous a fourni une image plus complète et plus solide de l'occupation de niche fonctionnelle dans les communautés végétales. De plus, nous avons réussi à relier le modèle de l'occupation de la niche fonctionnelle et le modèle d'équilibre dynamique et avons constaté que le filtrage concurrentiel était le processus dominant qui détermine le mode d'occupation de la niche fonctionnelle et la richesse des espèces le long du stress et de la perturbation des gradients. Au chapitre 4, nous fournissons un soutien empirique à une méthode d'ordination CSR calibrée globalement en montrant une relation entre l'abondance relative d'espèces en croissance dans les mésocosmes ayant différents niveaux de fertilité du sol et mortalité indépendante de la densité et leur classification CSR. Au chapitre 5, nous avons montré que la succession d'installations au cours de sept ans dans ces mésocosmes était plus déterministe d'un point de vue fonctionnel, mais plus historiquement contingent d'un point de vue taxonomique et que l'importance relative de la contingence historique a diminué à mesure que l'environnement devenait plus stressé ou perturbé. En conclusion, les structures de l'occupation de l'espace fonctionnel dans (le volume fonctionnel total, le chevauchement fonctionnel et le volume fonctionnel moyen, les Chapitre 2 et 3) ou entre les communautés locales (dissimilarité fonctionnelle, chapitre 5) sont déterministes plutôt que neutres (ou contingence historique ). Les espèces tolératrices de stress sont plus avantagées dans les mésocosmes moins fertiles tandis que les espèces rudérales sont plus avantagées dans les mésocosmes avec plus de mortalité indépendante de la densité.
Abstract : The patterns of niche occupancy within local communities, the spatial variability of biodiversity along environmental gradients of stress and disturbance, and the processes of plant succession are several fundamental topics in ecology. Recently, the trait-based approach has emerged as a promising way to understand the processes structuring plant communities and has even been proposed as a method to rebuild community ecology based on functional traits. Therefore, linking these fundamental themes through a functional lens should give us more insight into some basic questions in ecology and will be the main objective of my thesis. Generally, my PhD project is to investigate the structures of functional space occupancy along both spatial and temporal gradients. Specifically, the objective of Chapter 2 is to investigate the patterns of functional niche occupancy by calculating three key niche metrics (the total functional niche volume, the functional niche overlap and the average functional niche volume) from speciespoor communities to species-rich communities and to determine the main driver of the observed pattern of functional niche occupancy across plant communities worldwide. In Chapter 3, I aim to predict and explain the variation of species richness along gradients of stress and disturbance, by linking the dynamic equilibrium model and functional niche occupancy based on the framework developed in Chapter 2. The objective of Chapter 4 is to experimentally test the application of a globally calibrated CSR ordination method based on three leaf traits (leaf area, leaf dry matter content and specific leaf area) in local studies. Finally, the aim of Chapter 5 is to experimentally test the hypotheses reconciling the deterministic and historically contingent views of plant succession, by investigating the variation of taxonomic and functional dissimilarities between communities along gradients of stress and disturbance. The global study (Chapter 2) is based on a collection 21 trait datasets, spanning tropical to temperate biomes, and consisting of 313 plant communities representing different growth forms. The local studies (Chapter 3, 4 and 5) are based on the same experimental system consisting of 24 mesocosms experiencing different levels of stress and disturbance. The experiment started in 2009 with the same seed mixture of 30 herbaceous species broadcast over the 24 mesocosms and ended in 2016. We allowed natural colonization of seeds from the common soil seed bank and from the surroundings during the seven-year succession. Ten traits were measured on five individuals (sampled directly from the mesocosms) per species per mesocosms in 2014 (Chapter 3 and 4). Another set of traits (16 traits including some traits that were not able to measured directly in the mesocosms) were measured at the species level (species mean traits values) for the 34 most abundant species (some species disappeared in the mesocosms) over the seven years, by regrowing them separately for one growing season. In Chapter 2, we found communities were more functionally diverse (an increase in total functional volume) in species-rich communities, and species overlapped more within the community (an increase in functional overlap) but did not more finely divide the functional space (no decline in average functional volume). Moreover, habitat filtering is a widespread process driving the pattern of functional niche occupancy across plant communities. In Chapter 3, we found a similar pattern of functional niche occupancy on an experimental system with a constant community spatial size and trait-sampling effort, which together with Chapter 2 provided us a more comprehensive and robust picture of functional niche occupancy across plant communities. In addition, we succeeded in linking the pattern of functional niche occupancy and the dynamic equilibrium model and found that habitat filtering was the dominant process determining the pattern of functional niche occupancy and species richness along the gradients stress and disturbance. In Chapter 4, we provide empirical support for a globally calibrated CSR ordination method by showing a relationship between the relative abundance of species growing in mesocosms having different levels of soil fertility and density-independent mortality and their CSR classification. In Chapter 5, we showed that plant succession over seven years in these mesocosms was more deterministic from a functional perspective but more historically contingent from a taxonomic perspective, and that the relative importance of historical contingency decreased as the environment became more stressed or disturbed. In conclusion, the structures of functional space occupancy within (the total functional volume, the functional overlap and the average functional volume; Chapter 2 and 3) or between local communities (functional dissimilarity, Chapter 5) are deterministic rather than neutral (or historical contingency). Stress-tolerators were more favored in high stress communities, while ruderals are more favored in high disturbed mesocosms (Chapter 4).

The flood of 2000 caused extensive changes within the riparian landscape of the Sabie River, Kruger National Park (KNP). Changes within the riparian landscape and the removal of vegetation resulted in considerable changes in riparian vegetation characteristics. Open patches created by the flood served as a template for the establishment of new species and the regeneration of existing species, which consequently resulted in a patch mosaic. This memorable event encouraged an investigation into the response of the Sabie River ecosystem to the memorable Large Infrequent Disturbance (LID). Riparian ecosystems are driven by varying combinations of environmental factors, such as water availability, disturbance, herbivory, fire and river morphology. This complexity depicts unique vegetation structure and assemblages of associated plant species. The lack of sufficient knowledge on the role of riparian vegetation in the health assessment of surrounding ecosystems along semi-arid rivers prompted the establishment of the Kruger Rivers Post Flood Research Program (KRPFRP). Research conducted through this monitoring program four years after the 2000 flood, revealed no significant changes in the species composition, although the location and density of many common riparian species have been changed. There was a decrease in species density across the macro channel floor (MCF) and an increase in species density across the macro channel bank (MCB). Furthermore, it was reported that the flood altered the distribution of height classes across the macro channel. In general the riparian vegetation was shorter and bushier four years post-flood. These studies furthermore illustrated that the tree to shrub ratio did not change drastically from pre-flood conditions, although a decrease in the number of shrub individuals was reported. The research presented in this dissertation was designed to further explore changes in woody species composition and structure along the Sabie River, KNP at a post flood temporal interval, i.e. between the last survey in 2004 (by the KRPFRP) and 2010. For data compatibility, the sampling and analytical approach of this study conforms to the approach followed by the KRPFRP. Data were sampled within four preselected belt-transects that form part of the larger KRPFRP. All established woody individuals were counted and measured within each contiguous 10 m x 30 m plot within each of the four belt-transects. Log transformed species composition data were analysed through the application of the Bray Curtis dissimilarity index in combination with Ward’s method of clustering. Statistical significant differences between clusters were tested through the application of the Fisher’s exact relationship test. The MIXED Procedure or PROC MIXED model was used to investigate change within the vegetation structural data. Results obtained through the various analytical methods broadly support the findings of the KRPFRP. No significant change in woody species composition could be detected between 2004 and 2010. However, a change in the density (increase and decrease) of certain species across the MCB and MCF was revealed. Species richness and density increased significantly on the MCF oppose to small changes on the MCB. A significant increase in the total number of shrubs on the MCF contributed to an overall increase in woody density for the entire study area between 2004 and 2010. Shrubs therefore remained the most dominant growth form in both sampling years. Trees decreased across the MCB although the total number of established trees remained unchanged between 2004 and 2010. Riparian vegetation structure is directly linked to species assemblages, hence the continued dominance of shrub species along the Sabie River in the KNP The Sabie River riparian landscape is therefore still characterised by short and multi-stemmed woody individuals ten years after the LID.
Thesis (MSc (Environmental Sciences))--North-West University, Potchefstroom Campus, 2013

Biodiversity is determined by a myriad of complex processes acting at different scales. Given the current rates of biodiversity loss and change, it is of paramount importance that we improve our understanding of the underlying structure of ecological communities. In this thesis, I focused on Species Abundance Distributions (SAD), as a synthetic measure of biodiversity and community structure, and on Beta (β) diversity patterns, as a description of the spatial variation of species composition. I systematically assessed the effect of scale on both these patterns, analysing a broad range of community data, including different taxa and habitats, from the terrestrial, marine and freshwater realms. Knowledge of the scaling properties of abundance and compositional patterns must be fully integrated in biodiversity research if we are to understand biodiversity and the processes underpinning it, from local to global scales. SADs depict the relative abundance of the species present in a community. Although typically described by unimodal logseries or lognormal distributions, empirical SADs can also exhibit multiple modes. However, the existence of multiple modes in SADs has largely been overlooked, assumed to be due to sampling errors or a rare pattern. Thus, we do not know how prevalent multimodality is, nor do we have an understanding of the factors leading to this pattern. Here, I provided the first global empirical assessment of the prevalence of multimodality across a wide range of taxa, habitats and spatial extents. I employed an improved method combining two model selection tools, and (conservatively) estimated that ~15% of the communities were multimodal with strong support. Furthermore, I showed that the pattern is more common for communities at broader spatial scales and with greater taxonomic diversity (i.e. more phylogenetically diverse communities, since taxonomic diversity was measured as number of families). This suggests a link between multimodality and ecological heterogeneity, broadly defined to incorporate the spatial, environmental, taxonomic and functional variability of ecological systems. Empirical understanding of how spatial scale affects SAD shape is still lacking. Here, I established a gradient in spatial scale spanning several orders of magnitude by decomposing the total extent of several datasets into smaller subsets. I performed an exploratory analysis of how SAD shape is affected by area sampled, species richness, total abundance and taxonomic diversity. Clear shifts in SAD shape can provide information about relevant ecological and spatial mechanisms affecting community structure. There was a clear effect of area, species richness and taxonomic diversity in determining SAD shape, while total abundance did not exhibit any directional effect. The results supported the findings of the previous analysis, with a higher prevalence of multimodal SADs for larger areas and for more taxonomically diverse communities, while also suggesting that species spatial aggregation patterns can be linked to SAD shape. On the other hand, there was a systematic departure from the predictions of two important macroecological theories for SAD across scales, specifically regarding logseries distributions being selected only for smaller scales and when species richness and number of families were proportionally much smaller than the total extent. β diversity quantifies the variation in species composition between sites. Although a fundamental component of biodiversity, its spatial scaling properties are still poorly understood. Here, I tested if two conceptual types of β diversity showed systematic variation with scale, while also explicitly accounting for the two β diversity components, turnover and nestedness (species replacement vs species richness differences). I provided the first empirical analysis of β diversity scaling patterns for different taxa, revealing remarkably consistent scaling curves. Total β diversity and turnover exhibit a power law decay with log area, while nestedness is largely insensitive to scale changes. For the distance decay of similarity analysis, while area sampled affected the overall dissimilarity values, rates of similarity were consistent across large variations in sampled area. Finally, in both these analyses, turnover was the main contributor to compositional change. These results suggest that species are spatially aggregated across spatial scales (from local to regional scales), while also illustrating that substantial change in community structure might occur, despite species richness remaining relatively stable. This systematic and comprehensive analysis of SAD and community similarity patterns highlighted spatial scale, ecological heterogeneity and species spatial aggregation patterns as critical components underlying the results found. This work expanded the range of scales at which both theories deriving SAD and community similarity studies have been developed and tested (from local plots to continents). The results here showed strong departures from two important macroecological theories for SAD at different scales. In addition, the overall findings in this thesis clearly indicate that unified theories of biodiversity (or assuming a set of synthetic minimal assumptions) are unable to accommodate the variability in SADs shape across spatial scales reported here, and cannot fully reproduce community similarity patterns across scales. Incorporating more realistic assumptions, or imposing scale dependent assumptions, may prove to be a fruitful avenue for ecological research regarding the scaling properties of SAD and community similarity patterns. This will allow deriving new predictions and improving the ability of theoretical models to incorporate the variability in abundance and similarity patterns across scales.

We consider the problem of learning the parameters of a structured output prediction model, that is, learning to predict elements of a complex interdependent output space that correspond to a given input. Unlike many of the existing approaches, we focus on the weakly supervised setting, where most (or all) of the training samples have only been partially annotated. Given such a weakly supervised dataset, our goal is to estimate accurate parameters of the model by minimizing the regularized empirical risk, where the risk is measured by a user-specified loss function. This task has previously been addressed by the well-known latent support vector machine (latent SVM) framework. We argue that, while latent SVM offers a computational efficient solution to loss-based weakly supervised learning, it suffers from the following three drawbacks: (i) the optimization problem corresponding to latent SVM is a difference-of-convex program, which is non-convex, and hence susceptible to bad local minimum solutions; (ii) the prediction rule of latent SVM only relies on the most likely value of the latent variables, and not the uncertainty in the latent variable values; and (iii) the loss function used to measure the risk is restricted to be independent of true (unknown) value of the latent variables. We address the the aforementioned drawbacks using three novel contributions. First, inspired by human learning, we design an automatic self-paced learning algorithm for latent SVM, which builds on the intuition that the learner should be presented in the training samples in a meaningful order that facilitates learning: starting frome easy samples and gradually moving to harder samples. Our algorithm simultaneously selects the easy samples and updates the parameters at each iteration by solving a biconvex optimization problem. Second, we propose a new family of LVMs called max-margin min-entropy (M3E) models, which includes latent SVM as a special case. Given an input, an M3E model predicts the output with the smallest corresponding Renyi entropy of generalized distribution, which relies not only on the probability of the output but also the uncertainty of the latent variable values. Third, we propose a novel learning framework for learning with general loss functions that may depend on the latent variables. Specifically, our framework simultaneously estimates two distributions: (i) a conditional distribution to model the uncertainty of the latent variables for a given input-output pair; and (ii) a delta distribution to predict the output and the latent variables for a given input. During learning, we encourage agreement between the two distributions by minimizing a loss-based dissimilarity coefficient. We demonstrate the efficacy of our contributions on standard machine learning applications using publicly available datasets.

Un des objectifs de lanalyse de données symboliques est de permettre une meilleure modélisation des variations et des imprécisions des données réelles. Ces données expriment en effet, un niveau de connaissance plus élevé, la modélisation doit donc offrir un formalisme plus riche que dans le cadre de lanalyse de données classiques. Un ensemble dopérateurs de généralisation symbolique existent et permettent une synthèse et représentation des données par le formalisme des assertions, formalisme défini en analyse de données symboliques. Cette généralisation étant supervisée, est souvent sensible aux observations aberrantes. Lorsque les données que lon souhaite généraliser sont hétérogènes, certaines assertions incluent des observations virtuelles. Face à ce nouveau formalisme et donc cette extension dordre sémantique que lanalyse de données symbolique a apporté, une nouvelle approche de traitement et dinterprétation simpose. Notre objectif au cours de ce travail est daméliorer tout dabord cette généralisation et de proposer ensuite une méthode de traitement de ces données. Les contributions originales de cette thèse portent sur de nouvelles approches de représentation et de classification des données à structure complexe. Nous proposons donc une décomposition permettant daméliorer la généralisation tout en offrant le formalisme symbolique. Cette décomposition est basée sur un algorithme divisif de classification. Nous avons aussi proposé une méthode de généralisation symbolique non supervisée basée sur l'algorithme des cartes topologiques de Kohonen. L'avantage de cette méthode est de réduire les données d'une manière non supervisée et de modéliser les groupes homogènes obtenus par des données symboliques. Notre seconde contribution porte sur lélaboration dune méthode de classification traitant les données à structure complexe. Cette méthode est une adaptation de la version batch de lalgorithme des cartes topologiques de Kohonen aux tableaux de dissimilarités. En effet, seule la définition dune mesure de dissimilarité adéquate, est nécessaire pour le bon déroulement de la méthode.

Les travaux ont traité de l'ajustement des efforts de calcul vibro-acoustique dans le cadre d'une étude de qualité sonore sur un système vibro-acoustique dès le stade de la conception. Cette recherche d'ajustement a été appliquée au système vibro-acoustique constitué d'une plaque vibrante couplée à une cavité amortie. Pour un calcul fréquentiel, un premier pas dans ce sens demandait à ajuster deux paramètres de simulation importants : la fréquence maximale de calcul, puis le pas fréquentiel de calcul. En termes de contraintes perceptives, il a été cherché pour ces paramètres une valeur ajustée permettant de conserver les tendances qualitatives originellement établies à partir d'une évaluation auditive portant sur les sons réels rayonnés par la plaque dans la cavité. Les sons réels ont été collectés via l'expérimentation sur un banc d'essai constitué d'une plaque en acier couplée à une cavité parallélépipédique ; ces sons correspondaient au son rayonné par la plaque vibrante à l'intérieur de la cavité, enregistré dans diverses configurations structurales du système. Ces configurations impliquaient différentes combinaisons des modalités prises par les 3 paramètres structuraux variables retenus : l'épaisseur de la plaque, les conditions de serrage de la plaque et les propriétés en absorption de la cavité. Les sons enregistrés ont ensuite été soumis par paires à un jury d'auditeurs auxquels il était demandé d'évaluer leur dissimilarité et de donner un jugement de préférence. A partir de l'analyse des informations recueillies, des tendances qualitatives de référence ont été établies, en termes d'espace perceptif de dissimilarité, d'espace perceptif de préférence et de classement de préférence. Dès lors, en travaillant à partir des sons enregistrés, le processus d'ajustement a consisté, i) à déterminer, à partir d'un filtrage passe-bas appliqué aux sons, la fréquence de coupure ajustée qui permettait de préserver les tendances qualitatives de référence, ii) à rechercher le pas fréquentiel ajusté, à imposer aux spectres des sons avant synthèse, qui permettait de maintenir les tendances qualitatives précédemment établies.

Biodiversity is determined by a myriad of complex processes acting at different scales. Given the current rates of biodiversity loss and change, it is of paramount importance that we improve our understanding of the underlying structure of ecological communities. In this thesis, I focused on Species Abundance Distributions (SAD), as a synthetic measure of biodiversity and community structure, and on Beta (β) diversity patterns, as a description of the spatial variation of species composition. I systematically assessed the effect of scale on both these patterns, analysing a broad range of community data, including different taxa and habitats, from the terrestrial, marine and freshwater realms. Knowledge of the scaling properties of abundance and compositional patterns must be fully integrated in biodiversity research if we are to understand biodiversity and the processes underpinning it, from local to global scales. SADs depict the relative abundance of the species present in a community. Although typically described by unimodal logseries or lognormal distributions, empirical SADs can also exhibit multiple modes. However, the existence of multiple modes in SADs has largely been overlooked, assumed to be due to sampling errors or a rare pattern. Thus, we do not know how prevalent multimodality is, nor do we have an understanding of the factors leading to this pattern. Here, I provided the first global empirical assessment of the prevalence of multimodality across a wide range of taxa, habitats and spatial extents. I employed an improved method combining two model selection tools, and (conservatively) estimated that ~15% of the communities were multimodal with strong support. Furthermore, I showed that the pattern is more common for communities at broader spatial scales and with greater taxonomic diversity (i.e. more phylogenetically diverse communities, since taxonomic diversity was measured as number of families). This suggests a link between multimodality and ecological heterogeneity, broadly defined to incorporate the spatial, environmental, taxonomic and functional variability of ecological systems. Empirical understanding of how spatial scale affects SAD shape is still lacking. Here, I established a gradient in spatial scale spanning several orders of magnitude by decomposing the total extent of several datasets into smaller subsets. I performed an exploratory analysis of how SAD shape is affected by area sampled, species richness, total abundance and taxonomic diversity. Clear shifts in SAD shape can provide information about relevant ecological and spatial mechanisms affecting community structure. There was a clear effect of area, species richness and taxonomic diversity in determining SAD shape, while total abundance did not exhibit any directional effect. The results supported the findings of the previous analysis, with a higher prevalence of multimodal SADs for larger areas and for more taxonomically diverse communities, while also suggesting that species spatial aggregation patterns can be linked to SAD shape. On the other hand, there was a systematic departure from the predictions of two important macroecological theories for SAD across scales, specifically regarding logseries distributions being selected only for smaller scales and when species richness and number of families were proportionally much smaller than the total extent. β diversity quantifies the variation in species composition between sites. Although a fundamental component of biodiversity, its spatial scaling properties are still poorly understood. Here, I tested if two conceptual types of β diversity showed systematic variation with scale, while also explicitly accounting for the two β diversity components, turnover and nestedness (species replacement vs species richness differences). I provided the first empirical analysis of β diversity scaling patterns for different taxa, revealing remarkably consistent scaling curves. Total β diversity and turnover exhibit a power law decay with log area, while nestedness is largely insensitive to scale changes. For the distance decay of similarity analysis, while area sampled affected the overall dissimilarity values, rates of similarity were consistent across large variations in sampled area. Finally, in both these analyses, turnover was the main contributor to compositional change. These results suggest that species are spatially aggregated across spatial scales (from local to regional scales), while also illustrating that substantial change in community structure might occur, despite species richness remaining relatively stable. This systematic and comprehensive analysis of SAD and community similarity patterns highlighted spatial scale, ecological heterogeneity and species spatial aggregation patterns as critical components underlying the results found. This work expanded the range of scales at which both theories deriving SAD and community similarity studies have been developed and tested (from local plots to continents). The results here showed strong departures from two important macroecological theories for SAD at different scales. In addition, the overall findings in this thesis clearly indicate that unified theories of biodiversity (or assuming a set of synthetic minimal assumptions) are unable to accommodate the variability in SADs shape across spatial scales reported here, and cannot fully reproduce community similarity patterns across scales. Incorporating more realistic assumptions, or imposing scale dependent assumptions, may prove to be a fruitful avenue for ecological research regarding the scaling properties of SAD and community similarity patterns. This will allow deriving new predictions and improving the ability of theoretical models to incorporate the variability in abundance and similarity patterns across scales.
A biodiversidade é determinada por uma miríade de processos complexos que actuam a escalas diferentes. Face às actuais taxas de perda e alteração da biodiversidade, é vital melhorar a nossa compreensão da estrutura subjacente das comunidades ecológicas. Esta tese focou-se na análise de Species Abundance Distributions (SAD; Distribuição das abundâncias relativas das espécies), enquanto medida sintética de biodiversidade e da estrutura das comunidades, e de padrões de Beta (β) diversidade, enquanto medida de descrição da variação espacial na composição específica das comunidades. Os efeitos de escala nestes dois padrões de biodiversidade foram sistematicamente avaliados, analisando uma grande variedade de comunidades, incluindo diferentes taxa e habitats, dos reinos terrestre, marinho e água doce. O conhecimento sobre as propriedades de escala dos padrões de abundância e de composição específica das comunidades deve ser totalmente integrado na investigação da biodiversidade, no sentido de a podermos compreender melhor, bem como aos processos que a sustentam, desde escalas locais à escala global. As SADs descrevem a abundância relativa das espécies presentes numa comunidade. Apesar de serem tipicamente descritas por distribuições unimodais, como a logseries ou a lognormal, SADs empíricas podem também exibir várias modas. No entanto, a existência de múltiplas modas em SADs tem sido largamente ignorada, sendo normalmente assumida como um padrão raro ou atribuído a erros de amostragem. Desta forma, a frequência de multimodalidade em SADs é desconhecida, bem como os factores que podem levar à sua ocorrência. Nesta análise, efectuei a primeira avaliação empírica global da frequência de multimodalidade, analisando várias comunidades de differentes taxa, habitats e extensões espaciais. Usando um método melhorado que combina dois critérios de selecção de modelos, estimei (conservadoramente) que cerca de 15% das comunidades analisadas eram multimodais com grande suporte. Além disso, demonstrei que a multimodalidade é mais comum em comunidades com maior extensão espacial e com maior diversidade taxonómica (isto é, comunidades filogeneticamente mais diversas, uma vez que a diversidade taxonómica foi medida como o número de famílias). Estes resultados sugerem uma ligação entre SADs multimodais e heterogeneidade ecológica, aqui amplamente definida para incorporar a variabilidade espacial, ambiental, taxonómica e funcional dos sistemas ecológicos. Ainda não possuímos uma compreensão empírica de como a escala espacial afecta a forma das SADs. Nesta análise, estabeleci um gradiente de escala espacial abrangendo várias ordens de magnitude, começando por decompor a extensão espacial total de várias comunidades em secções menores. Este gradiente foi usado para realizar uma análise exploratória de como a forma das SADs é afectada pela área amostrada, riqueza específica, abundância total e diversidade taxonómica. Mudanças claras na forma das SADs podem fornecer informações sobre mecanismos ecológicos e espaciais relevantes que afectam a estrutura das comunidades. Esta análise demonstrou um efeito claro da área, riqueza específica e diversidade taxonómica na forma das SADs, enquanto que a abundância total não exibiu um efeito direccional. Estes resultados apoiam as conclusões da análise anterior, mostrando uma maior prevalência de SADs multimodais para áreas maiores e para comunidades mais diversas taxonomicamente. Adicionalmente, estes resultados sugerem que os padrões de agregação espacial das espécies influenciam a forma das SADs ao longo do gradiente espacial. Por outro lado, esta análise identificou diferenças sistemáticas relativamente às previsões de duas importantes teorias macroecológicas para as SAD a escalas diferentes, especificamente o facto de a logseries apenas ter sido seleccionada para escalas menores e quando a riqueza específica e o número de famílias eram proporcionalmente muito menores do que para a extensão total. A β diversidade quantifica a variação na composição específica entre locais. Apesar de ser um componente fundamental da biodiversidade, conhecimento sobre a variação das suas propriedades com a escala espacial ainda é escasso. Nesta análise, testei se dois tipos conceptuais de β diversidade apresentam variação sistemática com a escala, considerando também explicitamente os dois componentes de β diversidade: turnover e nestedness (aninhamento) – substituição de espécies vs diferenças na riqueza específica entre locais, respectivamente. Efectuei a primeira análise empírica de padrões de escala de β diversidade para diferentes taxa, revelando que as curvas de escala são notavelmente consistentes para as comunidades analisadas. A β diversidade total e a componente de turnover exibem um declínio segundo uma power law com o logaritmo da área, enquanto a componente de nestedness é basicamente insensível às mudanças de escala. Relativamente à análise do declínio da similaridade com a distância geográfica, enquanto a área amostrada afectou significativamente os valores de dissimilaridade total, as taxas de mudança na similaridade foram consistentes para grandes variações entre áreas amostradas. Finalmente, em ambas as análises, o turnover foi o principal contribuinte para as diferenças composicionais. Estes resultados sugerem que as espécies estão espacialmente agregadas ao longo das escalas espaciais analisadas (de locais a regionais). Adicionalmente, os resultados ilustram que mudanças substanciais na estrutura das comunidades podem ocorrer, apesar de a riqueza específica permanecer relativamente estável. A análise sistemática e abrangente de SADs e de padrões de similaridade nesta tese identificou a escala espacial, a heterogeneidade ecológica e padrões de agregação espacial das espécies como componentes críticos subjacentes aos resultados encontrados. Esta investigação expandiu as escalas às quais tanto teorias que derivam SAD, como estudos de similaridade têm sido desenvolvidos e testados (desde plots locais a continentes). Estes resultados identificaram claros desvios face a duas importantes teorias macroecológicas para SAD a diferentes escalas. Adicionalmente, os resultados gerais desta tese indicam claramente que teorias unificadas da biodiversidade (ou assumindo um conjunto mínimo de pressupostos sintéticos) não são capazes de, por um lado, acomodar a variabilidade na forma das SADs a escalas espaciais diferentes aqui reportada, e, por outro lado, reproduzir totalmente os padrões de similaridade a todas as escalas espaciais. A incorporação de pressupostos mais realistas, ou a imposição de pressupostos dependentes da escala, pode revelar-se uma linha de investigação produtiva para as propriedades de escala das SADs e de padrões de similaridade, permitindo derivar novas previsões e melhorar a capacidade dos modelos teóricos em incorporar a variabilidade nos padrões de abundância e de similaridade a várias escalas.
Apoio financeiro do POCTI no âmbito do III Quadro Comunitário de Apoio. Apoio financeiro da FCT e do FSE no âmbito do III Quadro Comunitário de Apoio
Programa Doutoral em Biologia e Ecologia das Alterações Globais

AbstractIn the last decades, the Paris metro area has experienced important structural changes linked to rising income inequality and a rapidly growing immigrant population. Using census data from 1990, 1999 to 2015, this chapter explores these transformations and how they have shaped trends in residential segregation. We find that the occupational structure of the area shifted upwards in the recent decade with a substantial increase in the share of the top occupational groups. This trend, however, did not primarily concern the immigrant population, which nonetheless experienced a growth in the middle class. These trends were further accompanied by an increase in income inequality driven by rising wages among the top 1% earners. Despite these changes, dissimilarity indexes between socioeconomic groups and between natives and immigrants have remained quite stable over the period. However, interaction indexes suggest that neighbourhoods are becoming more homogenous over time, both in terms of socioeconomic and ethnic diversity. Finally, the findings shed light on the correlation between socioeconomic and immigrant segregation. Socioeconomic disadvantage and the presence of immigrants within neighbourhoods, especially of non-European origin, are tightly correlated, and that correlation became stronger over time.

AbstractChina’s rapid economic growth since the early 1980s has been accompanied by a substantial increase in economic inequality. Economic restructuring, rural–urban migration, globalization and marketization have jointly led to a transformation of the socio-spatial structure of large Chinese cities. Although a handful of studies have examined the level and pattern of socioeconomic segregation in a particular Chinese city using neighbourhood-level census data from the year 2000, little research has been done to investigate in-depth changes in the level and pattern of segregation using more up to date and more geographically detailed data. This chapter aims to examine the levels, patterns and drivers of socioeconomic segregation in Shanghai, China, using neighbourhood-level and subdistrict-level data from the 2000 and 2010 decennial population census. This chapter uses the dissimilarity index to measure the overall level of socioeconomic segregation by occupation and household registration (hukou) status. Based on a location quotient and neighbourhood composition, it also illustrates the change in the spatial pattern of segregation. The chapter ends with a discussion on the possible drivers of segregation and policy suggestions to combat segregation in large Chinese cities.

AbstractSince the 2000s, Lima city shows important changes in its socio-spatial structure, decreasing the long-established opposition between the centre and the periphery, developing a more complex arrangement. Sustained national economic growth has allowed better socio-economic conditions in different areas of the city. However, high inequality still remains in the ways of production of urban space, which affects residential segregation. To identify possible changes in the segregation patterns of Metropolitan Lima, this study focuses on the spatial patterns of occupational groups, examining their causes and relation with income inequality. The analysis is based on the 1993 and 2007 census data, measuring residential segregation by the Dissimilarity Index, comparing with the Diversity Index. The results confirm trends towards increased segregation between occupational groups. Top occupational groups are concentrated in central areas, expanding into adjacent districts. Bottom occupational groups are over-represented in distant neighbourhoods. In-between, a new, more mixed, transitional zone has emerged in upgraded formerly low-income neighbourhoods. Areas of lower occupational diversity coincide with extreme income values, forming spaces of greater segregation. In the metropolitan centre–periphery pattern, the centre has expanded, while the periphery has been shifted to outer peripheral rings.

Patents contain valuable information for engineering design. However, the increasing number of annual patent publications makes it difficult for any individual designer to assimilate all up-to-date knowledge hidden in patent documents. In this paper, we proposed a computational approach to interpret design structure embedded in patent claims using pre-developed ontology libraries. The study combined natural language processing (NLP) techniques, text data-mining, ontological engineering, and our rule-based tree generation method. Data sources and adopted tools included online patent documents, knowledge gathered from engineering textbooks, WordNet, a part-of-speech tagger developed by the Stanford NLP group, and Graphviz. We showed that the framework proposed in the paper not only could help minimize manual work required for obtaining design structures but also enable automatic dissimilarity comparison between patents.

MD-805 is a potent thrombin-inhibitor having the structure of tri-pods; Arg skeletone, N-terminal side and C-terminal side. MD-805 showed weaker inhibitory activity to other enzymes than thrombin. In this report, to gather more detailed informations about the structural features of serine enzymes concerning the specificity, we experimentally examined the structure-activity relationship (SAR) of a number of arginine derivatives including MD-805 and theoretically generated a MD-805-trypsin complex model using the results of X-ray crystallography of MD-805 and BPTI-trypsin complex by calculation in principle to minimize van der Waals contacts, and thus we discussed to interpret SAR from the molecular level. SAR of C-terminal side of arginine derivatives was obtained with the inhibitory activity to trypsin, plasmin, and glandular kallikrein and compared with the previous results of thrombin, the followings being indicated: (1) The hydrophobic binding pocket (HBP), which was reported by us to be at least partly similar in stereogeometry between trypsin and thrombin, had the depth corresponded to the length of ethylpiperidine, (2) concerning the site (termed the P site) next to HBP, there were large differences in stereogeometry between trypsin and thrombin; the P site of trypsin could accept propyl and phenyl group attached to 4-position of piperidine, while that of thrombin was unable to accept them and (3) the P sites of plasmin and glandular kallikrein resembled that of trypsin in being able to accept phenyl group. MD-805-trypsin complex model supported the reasonable understanding that the stereogeometrical similarity in HBP between thrombin and trypsin was attributable to the high homology in amino acid sequences in Ser-195 loop and that the dissimilarity in the P sites between thrombin and the others was attributable to 9 amino acids insertion found only in thrombin (Loop B). Furthermore, many dansylarginine derivatives showed very strong inhibition for pseudocholinesterase, however, SAR for C-terminal side of these derivatives revealed the similarity and dissimilarity in HBP and the P site between pseudocholinesterase and the proteases described above.

In order to accurately estimate the fatigue life a floating structure, it is necessary to have a large set of discrete environmental conditions. If the damage to a structure largely stems from wave-induced forces, then the creation of a set of environmental conditions or ‘bins’ is trivial. However, when considering a floating platform supporting a wind turbine, it is necessary to consider not only the wave conditions, but also the wind conditions (and perhaps current, if possible). Thus, it is common to have greater than 5 dimensions in the timeseries (e.g., significant wave height, wave period, wave direction, wind speed, wind direction, etc). The creation of bins in two dimensions is quite easily solved by creating an arbitrary grid and taking the mean of all the observations which fall in a specific cell. In higher dimensions, a p-dimensional cell is not easily visualized and so the resulting set of bins cannot easily be graphically represented. In this paper, an iterative algorithm is developed to convert N observations, each with p-dimensions, into a set with M discrete bins, where M << N. The algorithm presented borrows heavily from the maximum dissimilarity algorithm used in a wide array of fields. The benefit of using this algorithm is that there is no ‘bias’ introduced by an initial grid from the user. That is, given a desired final number of clusters and a certain distance tolerance, a unique set of cluster exists for a given data set. Inherently, the algorithm selects a diverse array of observations, usually including extreme events or outliers, which may have undue impact on the fatigue life of a structure. Although the algorithm is computationally expensive O(N2M), reductions in computational cost are possible. Most importantly, the algorithm can be written in such a way that memory constraints are not an issue even for N = O(105). The clustering algorithm is described in both graphical and logical terms. A case study is presented, using publicly available data from the Netherlands Enterprise Agency. The data is visualized in two dimensions with the final number of bins equaling approximately 50, 100, 200, 500, 1000, and 2000 bins. These bins are compared with a previous algorithm from these authors. Various measures are presented to assess the fidelity of a set of bins with respect to the initial observations. Each set of bins are analyzed and it is clear the MDA-based algorithm outperforms the previous algorithm.