Rozprawy doktorskie na temat „SENSITIVE RULES”

Thesis (M.S.)--Worcester Polytechnic Institute.
Keywords: computation pushdown; XML; order-based Xquery processing; relational database; ordered SQL queries; data model mapping; XQuery; XML data mapping; SQL; XML algebra rewrite rules; XML document order. Includes bibliographical references (p. 64-67).

Strong objections appeared in the Czech Republic, especially in the automotive industry, against the EU-Korea Free Tra de Agreement right after its adoption in October 2009. There were fears that the Agreement would endanger Czech competitiveness in the European market while new opportunity for Czech exporters in the Korean market would be limited. The thesis aims to analyze the impacts of the Agreement on the Czech sensitive sectors as the author does not agree with the results of the very first Czech country specific analysis made by the Association for International Affairs (AMO) that neglects any important endangering effects of the Agreement. The crucial difference between this paper and the study made by AMO is that the author considers Czech competitiveness in the European market as a decisive criterion while AMO analysts analyze the impacts on the Czech market only. They do not consider Czech exports to the EU member states and potentially strengthening competition in the European market for Czech producers. In the thesis the sensitive sectors are defined using the data from the International Trade Centre (ITC), the Czech statistical office, and the UN COMTRADE. Hariss index is used to measure restrictiveness of the rules of origin for the very first time in the EU-Korean FTA context, the methodology of the European Commission is utilized to estimate potential savings in effect of duty drawback, and a qualitative analysis of the non-tariff barriers is applied. The author points at a decisive function of non tariff barriers that plays even more important role in terms of liberalization then tariff reduction itself. Regarding the automotive industry, the thesis shows that Czech competitiveness within heading 8703 (cars) will not be endangered in effect of the Agreement while there might be some difficulties within heading 8708 (parts & accessories for motor vehicles). The Czech Republic does not perform any revealed competitive advantage in trade of services. The arguments of Czech car makers against the Agreement about the trade within heading 8703 are not admitted. However further research is needed to analyze if the Agreement will have harmful effects on Czech competitiveness in the European market within heading 8708. The Agreement will not bring up any notable opportunity for Czech exporters in the Korean market.

Os cartões, sejam de crédito ou débito, são meios de pagamento altamente utilizados. Esse fato desperta o interesse de fraudadores. O mercado de cartões enxerga as fraudes como custos operacionais, que são repassados para os consumidores e para a sociedade em geral. Ainda, o alto volume de transações e a necessidade de combater as fraudes abrem espaço para a aplicação de técnicas de Aprendizagem de Máquina; entre elas, os classificadores. Um tipo de classificador largamente utilizado nesse domínio é o classificador baseado em regras. Entretanto, um ponto de atenção dessa categoria de classificadores é que, na prática, eles são altamente dependentes dos especialistas no domínio, ou seja, profissionais que detectam os padrões das transações fraudulentas, os transformam em regras e implementam essas regras nos sistemas de classificação. Ao reconhecer esse cenário, o objetivo desse trabalho é propor a uma arquitetura baseada em regras de associação e regressão logística - técnicas estudadas em Aprendizagem de Máquina - para minerar regras nos dados e produzir, como resultado, conjuntos de regras de detecção de transações fraudulentas e disponibilizá-los para os especialistas no domínio. Com isso, esses profissionais terão o auxílio dos computadores para descobrir e gerar as regras que embasam o classificador, diminuindo, então, a chance de haver padrões fraudulentos ainda não reconhecidos e tornando as atividades de gerar e manter as regras mais eficientes. Com a finalidade de testar a proposta, a parte experimental do trabalho contou com cerca de 7,7 milhões de transações reais de cartões fornecidas por uma empresa participante do mercado de cartões. A partir daí, dado que o classificador pode cometer erros (falso-positivo e falso-negativo), a técnica de análise sensível ao custo foi aplicada para que a maior parte desses erros tenha um menor custo. Além disso, após um longo trabalho de análise do banco de dados, 141 características foram combinadas para, com o uso do algoritmo FP-Growth, gerar 38.003 regras que, após um processo de filtragem e seleção, foram agrupadas em cinco conjuntos de regras, sendo que o maior deles tem 1.285 regras. Cada um desses cinco conjuntos foi submetido a uma modelagem de regressão logística para que suas regras fossem validadas e ponderadas por critérios estatísticos. Ao final do processo, as métricas de ajuste estatístico dos modelos revelaram conjuntos bem ajustados e os indicadores de desempenho dos classificadores também indicaram, num geral, poderes de classificação muito bons (AROC entre 0,788 e 0,820). Como conclusão, a aplicação combinada das técnicas estatísticas - análise sensível ao custo, regras de associação e regressão logística - se mostrou conceitual e teoricamente coesa e coerente. Por fim, o experimento e seus resultados demonstraram a viabilidade técnica e prática da proposta.
Credit and debit cards are two methods of payments highly utilized. This awakens the interest of fraudsters. Businesses see fraudulent transactions as operating costs, which are passed on to consumers. Thus, the high number of transactions and the necessity to combat fraud stimulate the use of machine learning algorithms; among them, rule-based classifiers. However, a weakness of these classifiers is that, in practice, they are highly dependent on professionals who detect patterns of fraudulent transactions, transform them into rules and implement these rules in the classifier. Knowing this scenario, the aim of this thesis is to propose an architecture based on association rules and logistic regression - techniques studied in Machine Learning - for mining rules on data and produce rule sets to detect fraudulent transactions and make them available to experts. As a result, these professionals will have the aid of computers to discover the rules that support the classifier, decreasing the chance of having non-discovered fraudulent patterns and increasing the efficiency of generate and maintain these rules. In order to test the proposal, the experimental part of the thesis has used almost 7.7 million transactions provided by a real company. Moreover, after a long process of analysis of the database, 141 characteristics were combined using the algorithm FP-Growth, generating 38,003 rules. After a process of filtering and selection, they were grouped into five sets of rules which the biggest one has 1,285 rules. Each of the five sets was subjected to logistic regression, so their rules have been validated and weighted by statistical criteria. At the end of the process, the goodness of fit tests were satisfied and the performance indicators have shown very good classification powers (AUC between 0.788 and 0.820). In conclusion, the combined application of statistical techniques - cost sensitive learning, association rules and logistic regression - proved being conceptually and theoretically cohesive and coherent. Finally, the experiment and its results have demonstrated the technical and practical feasibilities of the proposal.

Ce travail se propose d’aborder la question de l’éducation familiale, à l’heure où laconstitution d’un statut d’exception de l’enfant modifie les relations éducative et d’autoritéet rend plus incertains le jeu complexe entre individuel et collectif ainsi que la transmissiondes savoirs. Comment s'organisent les parents face cette question délicate de l'autorité,tiraillés entre les idéaux démocratiques et la composante contraignante irréductible de l'acteéducatif ? Quelle forme prend l’organisation familiale ? Vécu familial et façon de se mettreen règle à l'école sont-ils liés ?L’enjeu de ce travail est de mettre à jour les dynamiques qui ont participé de l’évolution del’éducation familiale par une étude des représentations de l’enfant et de l’enfance. Cetteétude vise ainsi à appréhender le déplacement des frontières d’éducation familiale tout enpermettant de reposer la question de l’autorité éducative par un croisement de la sociologiede la famille et la sociologie de l’école.L’étude porte sur les pratiques d’éducation familiale (organisation au quotidien avecl’enfant) et les valeurs éducatives promues par les parents. Elle révèle qu’elles nes’organisent plus de la même façon avec, sur le plan pratique, la mise en place d’une logiqued’hétérogénéité situationnelle, un délitement de l’ « éducation de classe » mais aussi lapermanence des idéologies sexistes. Sur le plan axiologique, apparaît l’homogénéisationprédite par Fize (1990) avec l’adhésion massive des parents aux valeurs promues par lesfamilles de milieu favorisé (épanouissement personnel de l’enfant).Enfin, avec l’étude de la sensibilité à l’ordre scolaire (façon de se comporter des élèves faceaux règles de vie scolaire) la thèse montre en quoi les multiples injonctions au recours audialogue et à l’explicitation des règles en éducation ne font que renforcer les difficultéséducatives. Elle révèle, au contraire, que l’essentiel de ce qui se joue dans l’éducation est del’ordre de l’indicible et que le sens des règles réside dans leur usage
This work is about family education, at the time when children’s status has changed. Thisnew status alters educational and authority relationships and questions the link betweenindividual and collectiv aspects of social life such as transmission. How do parents dealwith this sensitiv issue that is authority, torn between the democratic ideal and theconstraint that cannot be avoided in education ? How do they organize everyday life ? Isfamily organisation linked with children’s way to follow school rules ?Studying chilren and childhood representations, the issue of this work is to reveal thedynamics that contributed to family education evolution. Thus, the aim of this study is tounderstand the movement of family education while allowing to question educationnalauthority by combining family and school sociological approaches.The study focuses on parenting practices (parents’ daily organisation with their children)and educational values that parents promot. It reavels that they their organisation haschanged. Henceforth parenting practicises are structured by situational heterogeneity and« social class education » is less relevant to describe parenting practices unlike genderideologies. On the axiological level appears the educational homogenization that Fize(1990) predicted : parents massively join values promoted by privileged backgroundsfamilies (child’s personal development).Finally, with the notion of sensitivy to school form (students’ way to follow school rules),the thesis shows how, in education, dialogue and explication injonctions reinforceeducational difficulties. On the contrary, it reveals that most critical points in education areunspeakable and that the meaning of rules relies in their use

碩士
中原大學
資訊工程研究所
98
As the prevalent development of network technology, information sharing and communication is frequent in daily life. Although data mining techniques help people find the important rules among data, people also have to take the risk of sensitive information disclosed. In addition to hiding sensitive rules, recent researches also start discussing the reduction or avoidance of unexpected side effects, including the hiding of non-sensitive rules (lost rules) and the creation of non-existent rules (false rules). This thesis aims at a small amount of transaction modifications and proposes a method of rule hiding. The method can recover lost rules and keep sensitive rules hidden. Besides, we propose an efficient index structure for a quick retrieval of transactions during the hiding process. The experiments verify that our method can hide all sensitive rules and recover at most 40% lost rules. The index structure reduces about 85% retrieval time on average.

碩士
國立臺灣科技大學
資訊工程系
101
As the advancement of technologies as well as the intense competition of business, the issues of privacy have acquiring more attention. Mining association rule is the significant technique in data mining. However, it may cause some privacy problem in mining processes. Many researches, thus, start to hide sensitive association rules due to avoid the sensitive information exposed. However, the development of computers and Internet technologies is so fast that data are increasing successively. In addition, sensitive association rules will change with time and policy. These both are the challenges for protecting sensitive association rules. Most exist technologies of hiding sensitive association rules cannot handle dynamic data and sensitive rules effectively. For solving these problems, this paper proposed a framework to protect dynamic sensitive association rules in incremental environment, HSAi and HDSA. HSAi is the algorithm to protect sensitive association rule in incremental data and we design the strategy to select appropriate victim transactions and items to delete them in order to hide sensitive association rules. HDSA is the algorithm for protecting dynamic sensitive rules, including adding and deleting. The mean of the deleting sensitive rule is the association rule that hidden can show again in the mining result. The goals of HSAi and HDSA are not only protecting sensitive rules but also producing least side effect from released dataset. Experiment results represent that the framework situation of incremental data and dynamic sensitive rules both can cause least side effects and maintain a desirable quality of sanitized database as well.

碩士
國立東華大學
資訊工程學系
93
In the researches of data mining, discovering frequent patterns from huge amounts of data is one of the most studied problems. The frequent patterns mined form databases can bring the users many commercial benefits. However, some sensitive patterns with security concerned may cause a threat to privacy. We investigate to find an appropriate balance between a need for privacy and information discovery on frequent patterns. In this thesis, a novel method for modifying databases to hide sensitive patterns is proposed. By multiplying the original database and a sanitization matrix together, a sanitized database with privacy concerns is obtained. Additionally, two probability policies are introduced to against the recovery of sensitive patterns and reduce the probability of hiding non-sensitive patterns in the sanitized database. The complexity analysis of our sanitization process is proved and a set of experiments is also performed to show the benefit of our approach.

abstract: In the recent times, traffic congestion and motor accidents have been a major problem for transportation in major cities. Intelligent Transportation Systems has the potential to be an effective solution in order to tackle this issue. Connected Autonomous Vehicles can cooperate at intersections, ramp merging, lane change and other conflicting scenarios in order to resolve the conflicts and avoid collisions with other vehicles. A lot of works has been proposed for specific scenarios such as intersections, ramp merging or lane change which partially solve the conflict resolution problem. Also, one of the major issues in autonomous decision making - deadlocks have not been considered in some of the works. The existing works either do not consider deadlocks or lack a safety proof. This thesis proposes a cooperative driving solution that provides a complete navigation, conflict resolution and deadlock resolution for connected autonomous vehicles. A graph-based model is used to resolve the deadlocks between vehicles and the responsibility sensitive safety (RSS) rules have been used in order to ensure safety of the autonomous vehicles during conflict detection and resolution. This algorithm provides a complete navigation solution for an autonomous vehicle from its source to destination. The algorithm ensures that accidents do not occur even in the worst-case scenario and the decision making is deadlock free.
Dissertation/Thesis
Masters Thesis Computer Engineering 2020

Data mining process are followed an extensive undertaking of research and product improvement. This development started when enterprise material was first loaded on computers, continued with advancement in data access, and more recently, developed technologies that permit users to transport through their data in real time. APRIORI algorithm, a popular data mining technique and compared the performances of a linked list based implementation as a basis and a tries-based implementation on it for mining frequent item sequences in a transactional database. In this report, I study the data structure, implementation and algorithmic features mainly focusing on those that also arise in frequent item set mining. This algorithm has given us new capabilities to identify associations in large data sets. However, a fundamental issue, and still not adequately examined, is demand to balance the privacy of the revealed data with the legitimate needs of the data users. The rule is characterizing as sensitive if its disclosure threat is above a certain privacy threshold. Sometimes, sensitive rules should not disclose to the public, since among other things, they may use for inferring sensitive data, or they may offer enterprise competitors with an advantage. Therefore, next I worked with some association rule hiding algorithms and examined their performances to analyse their time complexity orderly and the affect that they have in the original database.

Залізняк О. А. Гендерно чутлива журналістика як засіб стереотипізації в інтернет-медіа : кваліфікаційна робота магістра спеціальності 061 "Журналістика" / наук. керівник Н. В. Романюк. Запоріжжя : ЗНУ, 2020. 70 с.
UA : Магістерська робота «Гендерно чутлива журналістика як засіб стереотипізації в інтернет-медіа» – основний текст – 70 сторінок. Для виконання дипломної роботи опрацьовано 70 джерел. Об’єктом дослідження є два сайти відомих українських ЗМІ: «ТСН» та «УНІАН». Предметом дослідження є особливості функціонування гендерно чутливої журналістики в інтернет-ресурсах «ТСН» і «УНІАН». Мета дослідження: виявлення особливостей функціонування гендерно чутливої журналістики як засобу стереотипізація. Методи дослідження: У роботі використано поєднання наукових підходів та принципів, які дали змогу комплексно та всебічно дослідити проблему стереотипізація в межах гендерно чутливої журналістики. Застосовуючи системний підхід, вдалося проаналізувати механізми формування гендерних стереотипів через онлайн-медіа. Найбільш вагому роль у дослідженні має група лінгвістичних методів, серед яких головну позицію займає контент-аналіз, який дозволив визначити тональність текстів через використання емоційно забарвленої лексики. Також було використано методи опису та систематизації, узагальнення, соціологічний метод, за допомогою яких виявлені та описані особливості функціонування гендерно чутливої журналістики, а також зроблені висновки щодо використання гендерно чутливої журналістики як засобу стереотипізація. Для реалізації поставленої мети необхідно виконати такі завдання: 1. Вивчити дефініції понять «гендерна чутливість», «стереотип», «стереотипізація», «інтернет-медіа», а також дати вичерпне тлумачення терміна «гендерно чутлива журналістика». 2. Окреслити механізми виникнення й існування стереотипів у ЗМІ. 3. Описати основні види стереотипів у сучасних ЗМІ. 5 4. Виявити специфіку поширення гендерно чутливої журналістики в українських інтернет-медіа на прикладі онлайн-ресурсів «ТСН» та «УНІАН». 5. Розглянути основні види стереотипів та можливі способи їх руйнування у ЗМІ. Методологічну і теоретичну основу дослідження складають наукові розвідки про психологію ЗМІ та теорію стереотипів таких авторів, як В. Ліппмана, Г. Мельник, М. Скорика, Р. Харріса. Також були вивчені й проаналізовані праці вітчизняних і зарубіжних учених: Н. Аджихіної, Н. Акопяна, С. Кузіна та В. Попова та ін. Наукова новизна одержаних результатів полягає у всебічному аналізі основних понять за темою: «гендерна чутливість», «гендерно чутлива журналістика», «стереотипізація», а також у спробі детально дослідити функціонування гендерно чутливої журналістики, виявити особливості формування та руйнування стереотипів засобами журналістики. Сфера застосування: матеріали проведеного дослідження можуть бути використані у професійній діяльності телевізійників, продюсерів, стереотипізація. Ця робота може стати у нагоді громадським організаціям, які опікуються додержанням гендерних норм, а також може бути використана під час вивчення курсів телевізійної журналістики, соціальної проблематики ЗМІ у вишах.
EN : Master's thesis "Gender-sensitive journalism as a means stereotyping in online media "- main text - 70 pages. For 70 sources were processed. The object of the study are two sites of well-known Ukrainian media: "TSN" and "UNIAN". The subject of the study is the peculiarities of gender functioning sensitive journalism in the Internet resources "TSN" and "UNIAN". The purpose of the study: to identify the features of gender functioning sensitive journalism as a means of stereotyping. Research methods: The work uses a combination of scientific approaches and principles that allowed for a comprehensive and comprehensive study the problem of stereotyping within gender-sensitive journalism. Applying a systems approach, it was possible to analyze the mechanisms formation of gender stereotypes through online media. The most important role in the study has a group of linguistic methods, among which the main position takes content analysis, which allowed to determine the tone of the texts through use of emotionally colored vocabulary. Methods were also used description and systematization, generalization, sociological method, using which identified and described the features of the functioning of gender-sensitive journalism, as well as conclusions on gender use sensitive journalism as a means of stereotyping. To achieve this goal it is necessary to perform the following tasks: 1. To study the definitions of "gender sensitivity", "stereotype", "Stereotyping", "online media", as well as give a comprehensive interpretation the term "gender sensitive journalism". 2. Outline the mechanisms of emergence and existence of stereotypes in the media. 3. Describe the main types of stereotypes in modern media. 5 4. Identify the specifics of the spread of gender-sensitive journalism Ukrainian Internet media on the example of online resources "TSN" and "UNIAN". 5. Consider the main types of stereotypes and possible ways to do them destruction in the media. The methodological and theoretical basis of the study are scientific intelligence on the psychology of the media and the theory of stereotypes by authors such as W. Lippman, G. Melnik, M. Skoryk, R. Harris. Also studied analyzed the works of domestic and foreign scientists: N. Adzhikhina, N. Hakobyan, S. Kuzin and V. Popov and others. The scientific novelty of the obtained results lies in a comprehensive analysis basic concepts on the topic: "gender sensitivity", "gender sensitive" journalism "," stereotyping ", as well as in an attempt to investigate in detail functioning of gender-sensitive journalism, identify features formation and destruction of stereotypes by means of journalism. Scope: the materials of the study can be used in the professional activities of television, producers, stereotyping. This work can be useful to public organizations that take care of compliance with gender norms, and can also be used under time to study courses in television journalism, social issues of the media in universities.

碩士
國立成功大學
資訊管理研究所
92
Neural network is the technology used in solving data mining problem. Usually , its result has higher accuracy and handle well in noise data ,also the network architecture can present the complex relation between attributes .However its learning result is a black box .User can’t easily understand neural network’s learning result which limit the application of neural networl. This research try to explain the learning result of neural network by extracting rules from trained neural network . The extracting architecture can be used in various kind of neural network model .We put the effect of misclassification cost into consideration during the rule extracting process .The architecture of our research is based on PRISM which is proposed by Cendrowska.We combined the modified PRISM model with Adacost ,Metacost and Information cost function in our model to make the whole extracting architecture can consider the effect of misclassification cost .The architecture proposed in the research will compare the number of extracting rule ,accuracy of ruleset and misclassification cost with REFNE which is proposed by Zhou. The comparison is based on UCI-ML dataset .

碩士
國立臺灣科技大學
資訊工程系
97
Association rule mining is one of the most important techniques for intelligent system design and has been widely applied in a vast number of real applications. However, when the database is too large, classical algorithms cannot reasonably complete in time. The sampling approach that processes a subset of the database is a viable alternative. However, using only a subset of the database makes it impossible to achieve perfectly accurate rule extraction because the statistical property of the sample is most certainly different from that of the original database. One approach to improve the accuracy is to remove “outliers” from an initial simple random sample to obtain a final sample whose statistical property is more similar to that of the original database. In this paper, we propose to apply data clustering techniques to the initial sample before outlier removal is performed on the resulting clusters, so that outliers are removed based on the statistical properties in individual clusters, instead of the initial sample as a whole. However, clustering transactional data with very high dimensions is a difficult problem by itself. We solve this problem by employing locality sensitive hashing (LSH) as a means for data clustering. LSH is used to divide the initial sample into multiple possibly overlapping set of clusters. we evaluate several algorithms based on this strategy, and find that our proposals exhibit better accuracy or execution time, or both, than previous algorithms

碩士
國立清華大學
資訊工程學系
91
As the growth of computer technology has been advanced, the amount of data has been increasing with an extremely fast rate. A variety of methods for knowledge discovery and data mining have been developed to help people digest the huge number of data. One of the popular data mining research issues is association rule mining. Based on the techniques for mining association rules, the correlations between data items can be identified. However, the misuses of these methods may bring undesired side effects to the people. Recently, researchers have made great efforts at hiding association rules. In this thesis, we develop a new approach that can hide the sensitive information without generating undesired side effects. Our approach consists of three steps corresponding to three possible problems. At first, we adopt the template concept to identify either the set of modifiable transactions or the set of probably affected association rules. For efficiency, we design indexing facilities for fast retrieval of the required information in the transaction database. Second, among the selected transactions for hiding sensitive rules, we further select the transactions that will not hide any of the non-sensitive rules. At the third step, we examine these selected transactions to avoid generating extra rules. Iteratively, sensitive rules can be hidden and the undesired side effects are avoided. In the experiments, we show the effectiveness of our approach according to the three conditions and analyze the performance of different methods for database modifications. Moreover, the results also show that our proposed approach has perfect scalability to the database size. Specifically, the time of the approach that considers all the three conditions is just a little bit slower than the time of the one that do not consider the two side effects.