Classificação de dados utilizando algoritmos genéticos e lógica difusa

Várias das técnicas tradicionais de Mineração de Dados têm sido aplicadas com êxito e outras esbarram em limitações, tanto no desempenho como na qualidade do conhecimento gerado. Pesquisas recentes têm demonstrado que as técnicas na área de IA, tais como Algoritmo Genético (AG) e Lógica Difusa (LD), podem ser utilizadas com sucesso. Nesta pesquisa o interesse é revisar algumas abordagens que utilizam AG em combinação com LD de forma híbrida para realizar busca em espaços grandes e complexos. Este trabalho apresenta o Algoritmo Genético (AG), utilizando Lógica Difusa, para a codificação, avaliação e reprodução dos cromossomos, buscando classificar dados através de regras extraídas de maneira automática com a evolução dos cromossomos. A Lógica Difusa é utilizada para deixar as regras mais claras e próximas da linguagem humana, utilizando representações lingüísticas para identificar dados contínuos.

Aplicação de redes NeuroFuzzy ao processamento de peças automotivas por meio de injeção de polímeros

The injection molding of automotive parts is a complex process due to the many non-linear and multivariable phenomena that occur simultaneously. Commercial software applications exist for modeling the parameters of polymer injection but can be prohibitively expensive. It is possible to identify these parameters analytically, but applying classical theories of transport phenomena requires accurate information about the injection machine, product geometry, and process parameters. However, neurofuzzy networks, which achieve a synergy by combining the learning capabilities of an artificial neural network with a fuzzy set's inference mechanism, have shown success in this field. The purpose of this paper was to use a multilayer perceptron artificial neural network and a radial basis function artificial neural network combined with fuzzy sets to produce an inference mechanism that could predict injection mold cycle times. The results confirmed neurofuzzy networks as an effective alternative to solving such problems.

Tecniche di elaborazione di dati macrosismici per la definizione di sorgenti sismogenetiche

La valutazione dell’intensità secondo una procedura formale trasparente, obiettiva e che permetta di ottenere valori numerici attraverso scelte e criteri rigorosi, rappresenta un passo ed un obiettivo per la trattazione e l’impiego delle informazioni macrosismiche. I dati macrosismici possono infatti avere importanti applicazioni per analisi sismotettoniche e per la stima della pericolosità sismica. Questa tesi ha affrontato il problema del formalismo della stima dell’intensità migliorando aspetti sia teorici che pratici attraverso tre passaggi fondamentali sviluppati in ambiente MS-Excel e Matlab: i) la raccolta e l’archiviazione del dataset macrosismico; ii), l’associazione (funzione di appartenenza o membership function) tra effetti e gradi di intensità della scala macrosismica attraverso i principi della logica dei fuzzy sets; iii) l’applicazione di algoritmi decisionali rigorosi ed obiettivi per la stima dell’intensità finale. L’intera procedura è stata applicata a sette terremoti italiani sfruttando varie possibilità, anche metodologiche, come la costruzione di funzioni di appartenenza combinando le informazioni macrosismiche di più terremoti: Monte Baldo (1876), Valle d’Illasi (1891), Marsica (1915), Santa Sofia (1918), Mugello (1919), Garfagnana (1920) e Irpinia (1930). I risultati ottenuti hanno fornito un buon accordo statistico con le intensità di un catalogo macrosismico di riferimento confermando la validità dell’intera metodologia. Le intensità ricavate sono state poi utilizzate per analisi sismotettoniche nelle aree dei terremoti studiati. I metodi di analisi statistica sui piani quotati (distribuzione geografica delle intensità assegnate) si sono rivelate in passato uno strumento potente per analisi e caratterizzazione sismotettonica, determinando i principali parametri (localizzazione epicentrale, lunghezza, larghezza, orientazione) della possibile sorgente sismogenica. Questa tesi ha implementato alcuni aspetti delle metodologie di analisi grazie a specifiche applicazioni sviluppate in Matlab che hanno permesso anche di stimare le incertezze associate ai parametri di sorgente, grazie a tecniche di ricampionamento statistico. Un’analisi sistematica per i terremoti studiati è stata portata avanti combinando i vari metodi per la stima dei parametri di sorgente con i piani quotati originali e ricalcolati attraverso le procedure decisionali fuzzy. I risultati ottenuti hanno consentito di valutare le caratteristiche delle possibili sorgenti e formulare ipotesi di natura sismotettonica che hanno avuto alcuni riscontri indiziali con dati di tipo geologico e geologico-strutturale. Alcuni eventi (1915, 1918, 1920) presentano una forte stabilità dei parametri calcolati (localizzazione epicentrale e geometria della possibile sorgente) con piccole incertezze associate. Altri eventi (1891, 1919 e 1930) hanno invece mostrato una maggiore variabilità sia nella localizzazione dell’epicentro che nella geometria delle box: per il primo evento ciò è probabilmente da mettere in relazione con la ridotta consistenza del dataset di intensità mentre per gli altri con la possibile molteplicità delle sorgenti sismogenetiche. Anche l’analisi bootstrap ha messo in evidenza, in alcuni casi, le possibili asimmetrie nelle distribuzioni di alcuni parametri (ad es. l’azimut della possibile struttura), che potrebbero suggerire meccanismi di rottura su più faglie distinte.

Knowledge representation through graphs

Due to the increasing amount of data, knowledge aggregation, representation and reasoning are highly important for companies. In this paper, knowledge aggregation is presented as the first step. In the sequel, successful knowledge representation, for instance through graphs, enables knowledge-based reasoning. There exist various forms of knowledge representation through graphs; some of which allow to handle uncertainty and imprecision by invoking the technology of fuzzy sets. The paper provides an overview of different types of graphs stressing their relationships and their essential features.

A first approach to an axiomatic model of multi-measures

We establish an axiomatic model of multi-measures, capturing some classes of measures studied in the fuzzy sets literature, where they are applied to only one or two arguments.

Contribución al estudio de la incompatibilidad y otras medidas relacionadas en la lógica borrosa

Resumen La investigación descrita en esta memoria se enmarca en el campo de la lógica borro¬sa. Más concretamente, en el estudio de la incompatibilidad, de la compatibilidad y de la suplementaridad en los conjuntos borrosos y en los de Atanassov. En este orden de ideas, en el primer capítulo, se construyen, tanto de forma directa como indirecta, funciones apropiadas para medir la incompatibilidad entre dos conjuntos borro-sos. Se formulan algunos axiomas para modelizar la continuidad de dichas funciones, y se determina si las medidas propuestas, y otras nuevas que se introducen, verifican algún tipo de continuidad. Finalmente, se establece la noción de conjuntos borrosos compatibles, se introducen axiomas para medir esta propiedad y se construyen algunas medidas de compa¬tibilidad. El segundo capítulo se dedica al estudio de la incompatibilidad y de la compatibilidad en el campo de los conjuntos de Atanassov. Así, en primer lugar, se presenta una definición axiomática de medida de incompatibilidad en este contexto. Después, se construyen medidas de incompatibilidad por medio de los mismos métodos usados en el caso borroso. Además, se formulan axiomas de continuidad y se determina el tipo de continuidad de las medidas propuestas. Finalmente, se sigue un camino similar al caso borroso para el estudio de la compatibilidad. En el tercer capítulo, después de abordar la antonimia de conjuntos borrosos y de conjuntos de Atanassov, se formalizan las nociones de conjuntos suplementarios en estos dos entornos y se presenta, en ambos casos, un método para obtener medidas de suplementaridad a partir de medidas de incompatibilidad vía antónimos. The research described in this report pertains to the field of fuzzy logic and specifically studies incompatibility, compatibility and supplementarity in fuzzy sets and Atanassov's fuzzy sets. As such is the case, Chapter 1 describes both the direct and indirect construction of appropriate functions for measuring incompatibility between two fuzzy sets. We formulate some axioms for modelling the continuity of functions and determine whether the proposed and other measures introduced satisfy any type of continuity. Chapter 2 focuses on the study of incompatibility and compatibility in the field of Ata¬nassov's fuzzy sets. First, we present an axiomatic definition of incompatibility measure in this field. Then, we use the same methods to construct incompatibility measures as in the fuzzy case. Additionally, we formulate continuity axioms and determine the type of conti¬nuity of the proposed measures. Finally, we take a similar approach as in the fuzzy case to the study of compatibility. After examining the antonymy of fuzzy sets and Atanassov's sets, Chapter 3 formalizes the notions of supplementary sets in these two domains, and, in both cases, presents a method for obtaining supplementarity measures from incompatibility measures via antonyms.

Paired structures in knowledge representation.

In this position paper we propose a consistent and unifying view to all those basic knowledge representation models that are based on the existence of two somehow opposite fuzzy concepts. A number of these basic models can be found in fuzzy logic and multi-valued logic literature. Here it is claimed that it is the semantic relationship between two paired concepts what determines the emergence of different types of neutrality, namely indeterminacy, ambivalence and conflict, widely used under different frameworks (possibly under different names). It will be shown the potential relevance of paired structures, generated from two paired concepts together with their associated neutrality, all of them to be modeled as fuzzy sets. In this way, paired structures can be viewed as a standard basic model from which different models arise. This unifying view should therefore allow a deeper analysis of the relationships between several existing knowledge representation formalisms, providing a basis from which more expressive models can be later developed.

Data mining and simulation: a grey relationship demonstration

Fuzzy data has grown to be an important factor in data mining. Whenever uncertainty exists, simulation can be used as a model. Simulation is very flexible, although it can involve significant levels of computation. This article discusses fuzzy decision-making using the grey related analysis method. Fuzzy models are expected to better reflect decision-making uncertainty, at some cost in accuracy relative to crisp models. Monte Carlo simulation is used to incorporate experimental levels of uncertainty into the data and to measure the impact of fuzzy decision tree models using categorical data. Results are compared with decision tree models based on crisp continuous data.

Development of a probabilistic graphical structure from a model of mental health clinical expertise

This thesis explores the process of developing a principled approach for translating a model of mental-health risk expertise into a probabilistic graphical structure. Probabilistic graphical structures can be a combination of graph and probability theory that provide numerous advantages when it comes to the representation of domains involving uncertainty, domains such as the mental health domain. In this thesis the advantages that probabilistic graphical structures offer in representing such domains is built on. The Galatean Risk Screening Tool (GRiST) is a psychological model for mental health risk assessment based on fuzzy sets. In this thesis the knowledge encapsulated in the psychological model was used to develop the structure of the probability graph by exploiting the semantics of the clinical expertise. This thesis describes how a chain graph can be developed from the psychological model to provide a probabilistic evaluation of risk that complements the one generated by GRiST’s clinical expertise by the decomposing of the GRiST knowledge structure in component parts, which were in turned mapped into equivalent probabilistic graphical structures such as Bayesian Belief Nets and Markov Random Fields to produce a composite chain graph that provides a probabilistic classification of risk expertise to complement the expert clinical judgements

Developing a probabilistic graphical structure from a model of mental-health clinical risk expertise

This paper explores the process of developing a principled approach for translating a model of mental-health risk expertise into a probabilistic graphical structure. The Galatean Risk Screening Tool [1] is a psychological model for mental health risk assessment based on fuzzy sets. This paper details how the knowledge encapsulated in the psychological model was used to develop the structure of the probability graph by exploiting the semantics of the clinical expertise. These semantics are formalised by a detailed specification for an XML structure used to represent the expertise. The component parts were then mapped to equivalent probabilistic graphical structures such as Bayesian Belief Nets and Markov Random Fields to produce a composite chain graph that provides a probabilistic classification of risk expertise to complement the expert clinical judgements. © Springer-Verlag 2010.

Task-based image annotation and retrieval

In order to address problems of information overload in digital imagery task domains we have developed an interactive approach to the capture and reuse of image context information. Our framework models different aspects of the relationship between images and domain tasks they support by monitoring the interactive manipulation and annotation of task-relevant imagery. The approach allows us to gauge a measure of a user's intentions as they complete goal-directed image tasks. As users analyze retrieved imagery their interactions are captured and an expert task context is dynamically constructed. This human expertise, proficiency, and knowledge can then be leveraged to support other users in carrying out similar domain tasks. We have applied our techniques to two multimedia retrieval applications for two different image domains, namely the geo-spatial and medical imagery domains. © Springer-Verlag Berlin Heidelberg 2007.

Methods of Adaptive Extraction and Analysis of Knowledge for Knowledge-base Construction and Fast Decision Making

An approach for knowledge extraction from the information arriving to the knowledge base input and also new knowledge distribution over knowledge subsets already present in the knowledge base is developed. It is also necessary to realize the knowledge transform into parameters (data) of the model for the following decision-making on the given subset. It is assumed to realize the decision-making with the fuzzy sets’ apparatus.

Qualitative Comparative Analysis as a Tool for Concept Clarification, Typology Building, and Contextualized Comparisons in Gender and Feminist Research

Qualitative Comparative Analysis (QCA) is a method for the systematic analysis of cases. A holistic view of cases and an approach to causality emphasizing complexity are some of its core features. Over the last decades, QCA has found application in many fields of the social sciences. In spite of this, its use in feminist research has been slower, and only recently QCA has been applied to topics related to social care, the political representation of women, and reproductive politics. In spite of the comparative turn in feminist studies, researchers still privilege qualitative methods, in particular case studies, and are often skeptical of quantitative techniques (Spierings 2012). These studies show that the meaning and measurement of many gender concepts differ across countries and that the factors leading to feminist success and failure are context specific. However, case study analyses struggle to systematically account for the ways in which these forces operate in different locations.

A two-step approach for the analysis of hybrids in comparative social policy analysis: a nuanced typology of childcare policies

Typologies have represented an important tool for the development of comparative social policy research and continue to be widely used in spite of growing criticism of their ability to capture the complexity of welfare states and their internal heterogeneity. In particular, debates have focused on the presence of hybrid cases and the existence of distinct cross-national pattern of variation across areas of social policy. There is growing awareness around these issues, but empirical research often still relies on methodologies aimed at classifying countries in a limited number of unambiguous types. This article proposes a two-step approach based on fuzzy-set-ideal-type analysis for the systematic analysis of hybrids at the level of both policies (step 1) and policy configurations or combinations of policies (step 2). This approach is demonstrated by using the case of childcare policies in European economies. In the first step, parental leave policies are analysed using three methods – direct, indirect, and combinatory – to identify and describe specific hybrid forms at the level of policy analysis. In the second step, the analysis focus on the relationship between parental leave and childcare services in order to develop an overall typology of childcare policies, which clearly shows that many countries display characteristics normally associated with different types (hybrids and. Therefore, this two-step approach enhances our ability to account and make sense of hybrid welfare forms produced from tensions and contradictions within and between policies.