Introdução: Retratamentos endodônticos apresentam maior complexidade técnica e piores prognósticos quando comparados ao tratamento endodôntico inicial. Nesse contexto, uma investigação mais detalhada em relação aos fatores que afetam a exiquibilidade de se obter qualidade técnica satisfatória e reparo periapical é necessária. Técnicas empregadas para mineração de dados são pouco exploradas na área da Odontologia, ainda que apresentem potencial em contribuir com a descoberta do conhecimento. No presente estudo, padrões e fatores de risco relacionados à qualidade técnica e ao reparo periapical de retratamentos endodônticos foram investigados. Árvores de decisão foram geradas, sendo essa técnica complementada pela análise estatística convencional. Metodologia: Este estudo observacional incluiu 321 indivíduos com indicação de retratamento endodôntico atendidos por alunos de especialização em Endodontia. Foram coletados dados demográficos, referentes a história médica, ao diagnóstico, ao tratamento e a controles pós-operatórios, os quais foram transferidos para uma base de dados eletrônica. Após o preparo e pré-processamento de dados, foram selecionadas 32 variáveis independentes e 2 variáveis dependentes, as quais compreenderam os desfechos qualidade técnica do retratamento e reparo periapical. Estatísticas descritivas foram conduzidas a fim de determinar a frequência de dados ausentes, a distribuição das variáveis categóricas e a média e desvio-padrão de variáveis numéricas. Foram geradas árvores de decisão para a determinação de padrões relacionados aos desfechos, através do software de mineração de dados Weka (Waikato Environment of Knowledge Analysis, University of Waikato, New Zealand). Análises estatísticas convencionais foram conduzidas com auxílio do Software SPSS (SPSS Inc., Chicago, IL, USA), a fim de determinar fatores que poderiam interferir nos referidos desfechos. Resultados: Após o retratamento endodôntico, qualidade técnica satisfatória e reparo periapical foram obtidos em 65,20% e em 80,50% dos casos, respectivamente. A qualidade técnica do retratamento endodôntico foi afetada por vários fatores de risco, incluindo curvatura radicular severa (p < 0,001) e alterações na morfologia do canal radicular (p = 0,002). As árvores de decisão sugeriram padrões que combinam a ocorrência simultânea de raízes retas e reabsorções radiculares apicais com resultados tecnicamente insatisfatórios. O diâmetro da lesão periapical (p = 0,018), o grupo dentário (p = 0,015) e a presença de reabsorções apicais (p = 0,024) apresentaram associação significativa com o insucesso de retratamentos endodônticos. A análise de mineração de dados sugeriu que lesões periapicais extensas e qualidade da obturação insatisfatória no tratamento endodôntico inicial, apresentam mecanismos de interação entre a infecção intracanal e a resposta do hospedeiro que não foram totalmente elucidados, sendo necessários estudos complementares. Conclusão: Qualidade técnica satisfatória é afetada por diversos fatores de risco, entre eles, a presença de curvaturas radiculares severas e alterações na morfologia do canal radicular. A localização dos acidentes de procedimento exerce influência na obtenção da qualidade técnica. Fatores como o diâmetro da lesão periapical, o grupo dentário e as reabsorções radiculares apicais mostraram-se significativamente associados ao insucesso de retratamentos endodônticos. / Introduction: Non-surgical root canal retreatment presents higher technical complexity and poor prognosis compared to primary endodontic treatment. Within this context, a more detaild investigation on the factors affecting the feasibilty of achieving technical quality and periapical healing in teeth presenting secondary root canal infection is needed. Data mining approach is still little explored in the dentistry field, regardless of its potential to contribute to knowledge discovery. In the present study decision trees were complemented by conventional statistical analysis aiming to investigate patterns and risk factors related to technical quality and healing outcomes in non-surgical root canal retreatment. Methods: This observational study included 321 consecutive patients presenting for non-surgical root canal retreatment. Patients were treated by postgraduate students, following standard protocols. Data concerning demographic, medical, diagnostic, treatment and follow-up variables were transferred to an eletronic chart database (ECD). After data preprocessing and preparation a total of 32 independent variables and 2 dependent variables were defined. Basic statistics were tabled and provided the frequency of missing values, the distribution of categorical attributes and the mean and standard deviation values of numeric attributes. Decision trees were generated to predict patterns related to technical quality (satisfactory/unsatisfactory) and periapical healing (healed /failure), using J48 classification algorithm in Weka data mining software (Waikato Environment of Knowledge Analysis, University of Waikato, New Zealand). Statistical tests were performed using SPSS software (SPSS Inc., Chicago, IL, USA). Univariate and multivariate analytic methods were used to determine factors affecting endodontic retreatment technical quality and periapical healing. Results: After endodontic retreatment, technical outcome was satisfactory in 65.20%, and periapical healing was observed in 80.50% of the cases. Technical quality of endodontic retreatment was affected by several risk factors, including severity of root curvature (p < 0.001) and altered root canal morphology (p = 0.002). The decision trees suggested that patterns that combine straight root curvature and apical root resorption may prevent satisfactory technical outcomes. Periapical lesion area (p = 0.018), tooth type (p = 0.015) and apical resorption (p = 0.024) were shown to be significantly associated with endodontic retreatment failure. Data mining analysis suggested that large periapical lesions, as well as poor root filling quality in the initial endodontic treatment, present mechanisms that are not fully understood with regards to the interaction between intracanal infection and host response, which should be further investigated. Conclusions: Technical quality of endodontic retreatment is affected by several risk factors, including severity of root curvature and altered root canal morphology. The occurence of procedure accidents is especially relevant in the apical third of the roots, affecting the technical quality. Periapical lesion area, tooth type and apical resorption were shown to be significantly associated with endodontic retreatment failure.

Data mining can be used in healthcare industry to “mine” clinical data to discover hidden information for intelligent and affective decision making. Discovery of hidden patterns and relationships often goes intact, yet advanced data mining techniques can be helpful as remedy to this scenario. This thesis mainly deals with Intelligent Prediction of Chronic Renal Disease (IPCRD). Data covers blood, urine test, and external symptoms applied to predict chronic renal disease. Data from the database is initially transformed to Weka (3.6) and Chi-Square method is used for features section. After normalizing data, three classifiers were applied and efficiency of output is evaluated. Mainly, three classifiers are analyzed: Decision Tree, Naïve Bayes, K-Nearest Neighbour algorithm. Results show that each technique has its unique strength in realizing the objectives of the defined mining goals. Efficiency of Decision Tree and KNN was almost same but Naïve Bayes proved a comparative edge over others. Further sensitivity and specificity tests are used as statistical measures to examine the performance of a binary classification. Sensitivity (also called recall rate in some fields) measures the proportion of actual positives which are correctly identified while Specificity measures the proportion of negatives which are correctly identified. CRISP-DM methodology is applied to build the mining models. It consists of six major phases: business understanding, data understanding, data preparation, modeling, evaluation, and deployment.

This thesis is concerned with the evaluation of real options whose value represents a certain flexibility of the firm to decide about company´s assets in the future. In addition to classic models which were developed for option rating, such as binomial and Black-Scholes model, which have advantages and disadvantages, there is introduced a possible combination of decision trees and simulation Monte Carlo which runs directly inside the tree. This combination can erase the disadvantages which these methods have when they are used separately for option evaluation. In this thesis there can be found an application example inspired by a real situation and there are described different possibilities of usage of the mentioned combination and there is also demonstrated an unambiguous advantage of this method and that is a bigger amount of information which is provided in comparison with standard models. It allows the company to access much more complex image of the investment. The result is also various option values according to the used technique.

Dentre as diversas tarefas em que os algoritmos evolutivos têm sido empregados, a indução de regras e de árvores de decisão tem se mostrado uma abordagem bastante atrativa em diversos domínios de aplicação. Algoritmos de indução de árvores de decisão representam uma das técnicas mais populares em problemas de classificação. Entretanto, os algoritmos tradicionais de indução apresentam algumas limitações, pois, geralmente, usam uma estratégia gulosa, top down e com particionamento recursivo para a construção das árvores. Esses fatores degradam a qualidade dos dados, os quais podem gerar regras estatisticamente não significativas. Este trabalho propõe o algoritmo LEGAL-Tree, uma nova abordagem baseada em algoritmos genéticos para indução de árvores de decisão. O algoritmo proposto visa evitar a estratégia gulosa e a convergência para ótimos locais. Para isso, esse algoritmo adota uma abordagem multi-objetiva lexicográfica. Nos experimentos realizados sobre bases de dados de diversos problemas de classificação, a função de fitness de LEGAL-Tree considera as duas medidas mais comuns para avaliação das árvores de decisão: acurácia e tamanho da árvore. Os resultados obtidos mostraram que LEGAL-Tree teve um desempenho equivalente ao algoritmo SimpleCart (implementação em Java do algoritmo CART) e superou o tradicional algoritmo J48 (implementação em Java do algoritmo C4.5), além de ter superado também o algoritmo evolutivo GALE. A principal contribuição de LEGAL-Tree não foi gerar árvores com maior acurácia preditiva, mas sim gerar árvores menores e, portanto, mais compreensíveis ao usuário do que as outras abordagens, mantendo a acurácia preditiva equivalente. Isso mostra que LEGAL-Tree obteve sucesso na otimização lexicográfica de seus objetivos, uma vez que a idéia era justamente dar preferência às árvores menores (em termos de número de nodos) quando houvesse equivalência de acurácia / Among the several tasks evolutionary algorithms have been successfully employed, the induction of classification rules and decision trees has been shown to be a relevant approach for several application domains. Decision tree induction algorithms represent one of the most popular techniques for dealing with classification problems. However, conventionally used decision trees induction algorithms present limitations due to the strategy they usually implement: recursive top-down data partitioning through a greedy split evaluation. The main problem with this strategy is quality loss during the partitioning process, which can lead to statistically insignificant rules. In this thesis we propose the LEGAL-Tree algorithm, a new GA-based algorithm for decision tree induction. The proposed algorithm aims to prevent the greedy strategy and to avoid converging to local optima. For such, it is based on a lexicographic multi-objective approach. In the experiments which were run in several classification problems, LEGAL-Tree\'s fitness function considers two of the most common measures to evaluate decision trees: accuracy and tree size. Results show that LEGAL-Tree performs similarly to SimpleCart (CART Java implementation) and outperforms J48 (C4.5 Java implementation) and the evolutionary algorithm GALE. LEGAL-Tree\'s main contribution is not to generate trees with the highest predictive accuracy possible, but to provide smaller (and thus more comprehensible) trees, keeping a competitive accuracy rate. LEGAL-Tree is able to provide both comprehensible and accurate trees, which shows that the lexicographic fitness evaluation is successful since its goal is to prioritize smaller trees (fewer number of nodes) when there is equivalency in terms of accuracy

Introdução: Retratamentos endodônticos apresentam maior complexidade técnica e piores prognósticos quando comparados ao tratamento endodôntico inicial. Nesse contexto, uma investigação mais detalhada em relação aos fatores que afetam a exiquibilidade de se obter qualidade técnica satisfatória e reparo periapical é necessária. Técnicas empregadas para mineração de dados são pouco exploradas na área da Odontologia, ainda que apresentem potencial em contribuir com a descoberta do conhecimento. No presente estudo, padrões e fatores de risco relacionados à qualidade técnica e ao reparo periapical de retratamentos endodônticos foram investigados. Árvores de decisão foram geradas, sendo essa técnica complementada pela análise estatística convencional. Metodologia: Este estudo observacional incluiu 321 indivíduos com indicação de retratamento endodôntico atendidos por alunos de especialização em Endodontia. Foram coletados dados demográficos, referentes a história médica, ao diagnóstico, ao tratamento e a controles pós-operatórios, os quais foram transferidos para uma base de dados eletrônica. Após o preparo e pré-processamento de dados, foram selecionadas 32 variáveis independentes e 2 variáveis dependentes, as quais compreenderam os desfechos qualidade técnica do retratamento e reparo periapical. Estatísticas descritivas foram conduzidas a fim de determinar a frequência de dados ausentes, a distribuição das variáveis categóricas e a média e desvio-padrão de variáveis numéricas. Foram geradas árvores de decisão para a determinação de padrões relacionados aos desfechos, através do software de mineração de dados Weka (Waikato Environment of Knowledge Analysis, University of Waikato, New Zealand). Análises estatísticas convencionais foram conduzidas com auxílio do Software SPSS (SPSS Inc., Chicago, IL, USA), a fim de determinar fatores que poderiam interferir nos referidos desfechos. Resultados: Após o retratamento endodôntico, qualidade técnica satisfatória e reparo periapical foram obtidos em 65,20% e em 80,50% dos casos, respectivamente. A qualidade técnica do retratamento endodôntico foi afetada por vários fatores de risco, incluindo curvatura radicular severa (p < 0,001) e alterações na morfologia do canal radicular (p = 0,002). As árvores de decisão sugeriram padrões que combinam a ocorrência simultânea de raízes retas e reabsorções radiculares apicais com resultados tecnicamente insatisfatórios. O diâmetro da lesão periapical (p = 0,018), o grupo dentário (p = 0,015) e a presença de reabsorções apicais (p = 0,024) apresentaram associação significativa com o insucesso de retratamentos endodônticos. A análise de mineração de dados sugeriu que lesões periapicais extensas e qualidade da obturação insatisfatória no tratamento endodôntico inicial, apresentam mecanismos de interação entre a infecção intracanal e a resposta do hospedeiro que não foram totalmente elucidados, sendo necessários estudos complementares. Conclusão: Qualidade técnica satisfatória é afetada por diversos fatores de risco, entre eles, a presença de curvaturas radiculares severas e alterações na morfologia do canal radicular. A localização dos acidentes de procedimento exerce influência na obtenção da qualidade técnica. Fatores como o diâmetro da lesão periapical, o grupo dentário e as reabsorções radiculares apicais mostraram-se significativamente associados ao insucesso de retratamentos endodônticos. / Introduction: Non-surgical root canal retreatment presents higher technical complexity and poor prognosis compared to primary endodontic treatment. Within this context, a more detaild investigation on the factors affecting the feasibilty of achieving technical quality and periapical healing in teeth presenting secondary root canal infection is needed. Data mining approach is still little explored in the dentistry field, regardless of its potential to contribute to knowledge discovery. In the present study decision trees were complemented by conventional statistical analysis aiming to investigate patterns and risk factors related to technical quality and healing outcomes in non-surgical root canal retreatment. Methods: This observational study included 321 consecutive patients presenting for non-surgical root canal retreatment. Patients were treated by postgraduate students, following standard protocols. Data concerning demographic, medical, diagnostic, treatment and follow-up variables were transferred to an eletronic chart database (ECD). After data preprocessing and preparation a total of 32 independent variables and 2 dependent variables were defined. Basic statistics were tabled and provided the frequency of missing values, the distribution of categorical attributes and the mean and standard deviation values of numeric attributes. Decision trees were generated to predict patterns related to technical quality (satisfactory/unsatisfactory) and periapical healing (healed /failure), using J48 classification algorithm in Weka data mining software (Waikato Environment of Knowledge Analysis, University of Waikato, New Zealand). Statistical tests were performed using SPSS software (SPSS Inc., Chicago, IL, USA). Univariate and multivariate analytic methods were used to determine factors affecting endodontic retreatment technical quality and periapical healing. Results: After endodontic retreatment, technical outcome was satisfactory in 65.20%, and periapical healing was observed in 80.50% of the cases. Technical quality of endodontic retreatment was affected by several risk factors, including severity of root curvature (p < 0.001) and altered root canal morphology (p = 0.002). The decision trees suggested that patterns that combine straight root curvature and apical root resorption may prevent satisfactory technical outcomes. Periapical lesion area (p = 0.018), tooth type (p = 0.015) and apical resorption (p = 0.024) were shown to be significantly associated with endodontic retreatment failure. Data mining analysis suggested that large periapical lesions, as well as poor root filling quality in the initial endodontic treatment, present mechanisms that are not fully understood with regards to the interaction between intracanal infection and host response, which should be further investigated. Conclusions: Technical quality of endodontic retreatment is affected by several risk factors, including severity of root curvature and altered root canal morphology. The occurence of procedure accidents is especially relevant in the apical third of the roots, affecting the technical quality. Periapical lesion area, tooth type and apical resorption were shown to be significantly associated with endodontic retreatment failure.

Machine Learning is growing in popularity and spreading across different fields for various applications. Due to this trend, machine learning algorithms use different hardware platforms and are being experimented to obtain high test accuracy and throughput. FPGAs are well-suited hardware platform for machine learning because of its re-programmability and lower power consumption. Programming using FPGAs for machine learning algorithms requires substantial engineering time and effort compared to software implementation. We propose a software assisted design flow to program FPGA for machine learning algorithms using our hardware library. The hardware library is highly parameterized and it accommodates Tree Classifiers. As of now, our library consists of the components required to implement decision trees and random forests. The whole automation is wrapped around using a python script which takes you from the first step of having a dataset and design choices to the last step of having a hardware descriptive code for the trained machine learning model.

We study the depth of decision trees for diagnosis of constant faults in read-once contact networks over finite bases. This includes diagnosis of 0-1 faults, 0 faults and 1 faults. For any finite basis, we prove a linear upper bound on the minimum depth of decision tree for diagnosis of constant faults depending on the number of edges in a contact network over that basis. Also, we obtain asymptotic bounds on the depth of decision trees for diagnosis of each type of constant faults depending on the number of edges in contact networks in the worst case per basis. We study the set of indecomposable contact networks with up to 10 edges and obtain sharp coefficients for the linear upper bound for diagnosis of constant faults in contact networks over bases of these indecomposable contact networks. We use a set of algorithms, including one that we create, to obtain the sharp coefficients.

Many analytics problems in Operations Research and the Management Sciences can be framed as decision-making problems containing uncertain input parameters to be estimated from data. For example, inventory optimization problems often require forecasts of future demand, and product recommendation systems (e.g., movies, sporting goods) depend on models for predicting customer responses to the feasible recommendations. Therefore, a question central to many analytics problems is how to optimally build models from data which estimate the uncertain inputs for the decision problems of interest. We argue that most common approaches for this task either (a) focus on the wrong objectives in training the models for the decision problem, or (b) focus on the right objectives but only study how to do so with prohibitively simple machine learning models (e.g. linear and logistic regression). In this work, we study how to train decision tree models for predicting uncertain parameters for analytical decision-making problems. Unlike other machine learning models such as linear and logistic regression, decision trees are both nonparameteric and interpretable, allowing them the capability of modeling highly complex relationships between data and predictions while also being easily visualized and interpreted. We propose tractable algorithms for decision tree training in the context of three problem domains relevant to Operations Research. First, we study how to train decision trees for delivering real-time personalized recommendations of products in settings where little prior data is available for training purposes. This problem is known in the literature as the contextual bandit problem and requires careful navigation of the so-called "exploration-exploitation trade-off" in utilizing the decision tree models. Second, we propose a new framework which we call Market Segmentation Trees (MSTs) for training decision tree models for the purposes of market segmentation and personalization. We explore several applications of MSTs relevant to personalized advertising, including recommending hotels to Expedia users as a function of their search queries and segmenting ad auctions according to the distribution of bids that they receive. Finally, we propose a general framework for training decision tree models for uncertain optimization problems which we call "SPO Trees" (SPOTs). In contrast to the typical objective of maximizing predictive accuracy, the SPOT framework trains decision trees to maximize the quality of the solutions found in the uncertain optimization problem, therefore yielding better decisions in several analytics problems of interest.

A large volume of data has become commonplace in many domains these days. Machine learning algorithms can be trained to look for any useful hidden patterns in such data. Sometimes, these big data might need to be summarized to make them into a manageable size, for example by using histograms, for various reasons. Traditionally, machine learning algorithms can be trained on data expressed as real numbers and/or categories but not on a complex structure such as histogram. Since machine learning algorithms that can learn from data with histograms have not been explored to a major extent, this thesis intends to further explore this domain. This thesis has been limited to classification algorithms, tree-based classifiers such as decision trees, and random forest in particular. Decision trees are one of the simplest and most intuitive algorithms to train. A single decision tree might not be the best algorithm in term of its predictive performance, but it can be largely enhanced by considering an ensemble of many diverse trees as a random forest. This is the reason why both algorithms were considered. So, the objective of this thesis is to investigate how one can adapt these algorithms to make them learn better on histogram data. Our proposed approach considers the use of multiple bins of a histogram simultaneously to split a node during the tree induction process. Treating bins simultaneously is expected to capture dependencies among them, which could be useful. Experimental evaluation of the proposed approaches was carried out by comparing them with the standard approach of growing a tree where a single bin is used to split a node. Accuracy and the area under the receiver operating characteristic (ROC) curve (AUC) metrics along with the average time taken to train a model were used for comparison. For experimental purposes, real-world data from a large fleet of heavy duty trucks were used to build a component-failure prediction model. These data contain information about the operation of trucks over the years, where most operational features are summarized as histograms. Experiments were performed further on the synthetically generated dataset. From the results of the experiments, it was observed that the proposed approach outperforms the standard approach in performance and compactness of the model but lags behind in terms of training time. This thesis was motivated by a real-life problem encountered in the operation of heavy duty trucks in the automotive industry while building a data driven failure-prediction model. So, all the details about collecting and cleansing the data and the challenges encountered while making the data ready for training the algorithm have been presented in detail.

C4.5 is a decision tree learning algorithm that was developed by Ross Quinlan based on his earlier algorithm ID3. C4.5 is one of the most popular algorithms used to solve classification problems. Classification problems are problems of interest in a variety of disciplines. C4.5 is a supervised learning algorithm which uses a set of training patterns to build a decision tree. The algorithm uses the patterns and analyzes their individual attributes to partition the pattern data. The popularity of C4.5 stems from the fact that it can handle both continuous and categorical attributes, and it can deal with missing attribute values, while at the same time providing an easy interpretation for the answers that it produces. There are two objectives of this thesis. The first is to implement C4.5 in C++ within a generic architecture to allow for additional modules to be added. The second is to use this generic architecture to implement an innovative post induction phase which adjusts splits to minimize the error of the C4.5 tree. The C4.5 code and the post induction phase will be compiled into a MEX DLL for use as functions within MATLAB. Experimentation is performed using MATLAB to verify the advantages of this post induction phase.