INF - Instituto de Informática

URI Permanente desta comunidade

http://200.137.215.59/tede/handle/tde/145

Navegar

Agora exibindo 1 - 15 de 15

Planejamentos combinatórios construindo sistemas triplos de steiner
(Universidade Federal de Goiás, 2011-08-26) Barbosa, Enio Perez Rodrigues; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610
Intuitively, the basic idea of Design Theory consists of a way to select subsets, also called blocks, of a finite set, so that some properties are satisfied. The more general case are the blocks designs. A PBD is an ordered pair (S;B), where S is a finite set of symbols, and B is a collection of subsets of S called blocks, such that each pair of distinct elements of S occur together in exactly one block of B. A Steiner Triple System is a particular case of a PBD, where every block has size only 3, being called triples. The main focus is in building technology systems. By resolvability is discussed as a Steiner Triple Systems is resolvable, and when it is not resolvable. This theory has several applications, eg, embeddings and even problems related to computational complexity.
Sobre grafos com r tamanhos diferentes de conjuntos independentes maximais e algumas extensões
(Universidade Federal de Goiás, 2014-10-01) Cappelle, Márcia Rodrigues; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Abreu, Nair Maria Maia de; Santos, José Plínio de Oliveira; Longo, Humberto José; Silva, Hebert Coelho da
In this thesis, we present some results concerning about the sizes of maximal independent sets in graphs. We prove that for integers r and D with r 2 and D 3, there are only finitely many connected graphs of minimum degree at least 2, maximum degree at most D, and girth at least 7 that have maximal independent sets of at most r different sizes. Furthermore, we prove several results restricting the degrees of such graphs. These contributions generalize known results on well-covered graphs. We study the structure and recognition of the well-covered graphs G with order n(G) without an isolated vertex that have independence number n(G)􀀀k 2 for some non-negative integer k. For k = 1, we give a complete structural description of these graphs, and for a general but fixed k, we describe a polynomial time recognition algorithm. We consider graphs G without an isolated vertex for which the independence number a(G) and the independent domination number i(G) satisfy a(G) 􀀀 i(G) k for some non-negative integer k. We obtain a upper bound on the independence number in these graphs. We present a polynomial algorithm to recognize some complementary products, which includes all complementary prisms. Also, we present results on well-covered complementary prisms. We show that if G is not well-covered and its complementary prism is well-covered, then G has only two consecutive sizes of maximal independent sets. We present an upper bound for the quantity of sizes of maximal independent sets in complementary prisms and other wellcovered concerning results. We present a lower bound for the quantity of different sizes of maximal independent sets in Cartesian products of paths and cycles.
Mineração de dados para classificação e caracterização de alguns vinhos Vitis Vinífera da América do Sul
(Universidade Federal de Goiás, 2016-12-21) Costa, Nattane Luíza da; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Leitão Júnior, Plínio de Sá; Guimarães, Marco Paulo
One concern regarding the production and marketing of wines is to ensure that the product is not adulterated in relation to the origin and type of grape used in its production. This is due to the high cost involved in production and due to interest of consumers in obtaining legitimate products. In this context, the techniques of data mining allow us to verify the relationship between the chemical properties of wines and their label regarding origin or type of grape. This study presents a method for classification and characterization of wines with the application of data mining to the chemical properties that describe the functionality of wines. Five applications were carried out involving Cabernet Sauvignon, Carménère, Syrah, Tannat and Merlot varieties produced in Argentina, Brazil, Chile and Uruguay: the classification of Cabernet Sauvignon according to geographic region of production, Brazil and Chile; the classification of Tannat wines from the southern regions of Uruguay and southern Brazil, regions in close proximity and relevant to the production of Tannat wines; the classification of Syrah wines from Argentina and Chile, which are close regions and have a significant production in the countries covered; the classification of Merlot wines associated with the four countries to draw a profile of the relevant variables for the classification of wines for each set of two countries; and the classification of wines of the Chilean Carménère and Merlot varieties, which aim to investigate a profile of discrimination between varieties. The results obtained in all applications are promising, with a high predictive performance of 88%. The combination of variable selection associated with the classifiers Support Vector Machines and Artificial Neural Networks made it possible to define classification models capable of predicting new samples in addition to identifying groups of variables responsible for the classification.
Uso e estabilidade de seletores de variáveis baseados nos pesos de conexão de redes neurais artificiais
(Universidade Federal de Goiás, 2021-03-19) Costa, Nattane Luíza da; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Lima, Márcio Dias de; Lins, Isis Didier; Costa, Ronaldo Martins da; Leitão Júnior, Plínio de Sá
Artificial Neural Networks (ANN) are machine learning models used to solve problems in several research fields. Although, ANNs are often considered “black boxes”, which means that these models cannot be interpreted, as they do not provide explanatory information. Connection Weight Based Feature Selectors (WBFS) have been proposed to extract knowledge from ANNs. Most of studies that have been using these algorithms are based on just one ANN model. However, there are variations in the ANN connection weight values due to the initialization and training, and consequently, leading to variations in the importance ranking generated by a WBFS. In this context, this thesis presents a study about the WBFS. First, a new voting approach is proposed to assess the stability of the WBFS, i.e, the variation in the result of the WBFS. Then, we evaluated the stability of the algorithms based on multilayer perceptron (MLP) and extreme learning machines (ELM). Furthermore, an improvement is proposed in the algorithms of Garson, Olden, and Yoon, combining them with the feature selector ReliefF. The new algorithms are called FSGR, FSOR, and FSYR. The experiments were performed based on 28 MLP architectures, 16 ELM architectures, and 16 data sets from the UCI Machine Learning Repository. The results show that there is a significant difference in WBFS stability depending on the training parameters of the ANNs and depending on the WBFS used. In addition, the proposed algorithms proved to be more effective than the classic algorithms. As far as we know, this study was the first attempt to measure the stability of WBFS, to investigate the effects of different ANN training parameters on the stability of WBFS, and the first to propose a combination of WBFS with another feature selector. Besides, the results provide information about the benefits and limitations of WBFS and represent a starting point for improving the stability of these algorithms.
Alianças defensivas em grafos
(Universidade Federal de Goiás, 2010-03-26) Dias, Elisângela Silva; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Martins, Wellington Santos; Tronto, Íris Fabiana de Barcelos
A defensive alliance in graph G = (V,E) is a set of vertices S ⊆V satisfying the condition that every vertex v ∈ S has at most one more neighbor in V − S than S. Due to this type of alliance, the vertices in S together defend themselves to the vertices in V − S. This dissertation introduces the basic concepts for the understanding of alliances in graphs, along with a variety of alliances and their numbers and provides some mathematical properties for these alliances, focusing mainly on defensive alliances in graphs. It shows theorems, corollaries, lemmas, propositions and observations with appropriate proofs with respect to the minimum degree of a graph G δ(G), the maximum degree ∆(G), the algebraic connectivity µ, the total dominanting set γt(G), the eccentricity, the edge connectivity λ(G), the chromatic number χ(G), the (vertex) independence number β0(G), the vertex connectivity κ(G), the order of the largest clique ω(G) and the domination number γ(G). It also shows a generalization of defensive alliances, called defensive k alliance, and the definition and properties of a security set in G. A secure set S ⊆ V of graph G = (V,E) is a set whose every nonempty subset can be successfully defended of an attack, under appropriate definitions of “attack” and “defence”.
Alianças defensivas em grafos
(Universidade Federal de Goiás, 2010-03-26) Dias, Elisângela Silva; Barbosa, Rommel Melgaço; http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K4786013D6; Martins, Wellington Santos; http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K4782112U1; Tronto, Íris Fabiana de Barcelos; http://buscatextual.cnpq.br/buscatextual/visualizacv.do?id=K4707515J4
A defensive alliance in graph G = (V;E) is a set of vertices S V satisfying the condition that every vertex v 2 S has at most one more neighbor in V 􀀀S than S. Due to this type of alliance, the vertices in S together defend themselves to the vertices in V 􀀀S. This dissertation introduces the basic concepts for the understanding of alliances in graphs, along with a variety of alliances and their numbers and provides some mathematical properties for these alliances, focusing mainly on defensive alliances in graphs. It shows theorems, corollaries, lemmas, propositions and observations with appropriate proofs with respect to the minimum degree of a graph G d(G), the maximum degree D(G), the algebraic connectivity μ, the total dominanting set gt(G), the eccentricity, the edge connectivity l(G), the chromatic number c(G), the (vertex) independence number b0(G), the vertex connectivity k(G), the order of the largest clique w(G) and the domination number g(G). It also shows a generalization of defensive alliances, called defensive kalliance, and the definition and properties of a security set in G. A secure set S V of graph G = (V;E) is a set whose every nonempty subset can be successfully defended of an attack, under appropriate definitions of “attack” and “defence”.
Sobre convexidade em prismas complementares
(Universidade Federal de Goiás, 2015-04-10) Duarte, Márcio Antônio; Szwarcfiter, Jayme L.; http://lattes.cnpq.br/2002515486942024; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Yanasse, Horacio Hideki; Oliveira, Carla Silva; Coelho, Erika Morais Martins; Silva, Hebert Coelho da
In this work, we present some related results, especially the properties algoritimics and of complexity of a product of graphs called complementary prism. Answering some questions left open by Haynes, Slater and van der Merwe, we show that the problem of click, independent set and k-dominant set is NP-Complete for complementary prisms in general. Furthermore, we show NP-completeness results regarding the calculation of some parameters of the P3-convexity for the complementary prism graphs in general, as the P3-geodetic number, P3-hull number and P3-Carathéodory number. We show that the calculation of P3-geodetic number is NP-complete for complementary prism graphs in general. As for the P3-hull number, we can show that the same can be efficiently computed in polynomial time. For the P3-Carathéodory number, we show that it is NPcomplete complementary to prisms bipartite graphs, but for trees, this may be calculated in polynomial time and, for class of cografos, calculating the P3-Carathéodory number of complementary prism of these is 3. We also found a relationship between the cardinality Carathéodory set of a graph and a any Carathéodory set of complementary prism. Finally, we established an upper limit calculation the parameters: geodetic number, hull number and Carathéodory number to operations complementary prism of path, cycles and complete graphs considering the convexities P3 and geodesic.
Mínimos quadrados para problemas de múltiplas classes envolvendo twin support vector machine e aplicações de mineração de dados
(Universidade Federal de Goiás, 2018-12-07) Lima, Márcio Dias de; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Santos, Helton Saulo Bezerra dos; Lozano, Kátia Kelvis Cassiano; Costa, Ronaldo Martins da; Rosa, Thierson Couto
Data mining is an emerging area due to the increasing amount of data available in a variety of fields. In this context twin support vector machine (TWSVM) has attracted the attention of several researchers. In this thesis, we developed a feature selector algorithm and an algorithm for multi-class problems based on TWSVM. This learning algorithm with ternary outputs {- 1,0,+1 } is based on the Vapnik support vector theory, and evaluates all training samples with a 1-×-1-×-rest structure during the decomposition phase. One of the main advantages of the proposed algorithm is the use of the least squares version for multi-class problems, where it is necessary to solve two systems of linear equations instead of two quadratic programming problems in TWSVM. We also implemented the principle of minimization of structural risk in order to improve the generalizability. The Sherman-Morisson-Woodbury formula is applied to reduce the complexity of the non-linear formulation of the algorithm. We also apply data mining techniques that combine the use of analytical technique with data mining algorithms in the classification of several samples. The developed framework could be an excellent tool for detecting different types of fraud, verifying if products were grown in organic or conventional systems, as well as tracing the region of origin of wine made from a given type of grape.
Mineração de dados para o reconhecimento da origem e do tipo de alimentos e outras substâncias com base em sua composição química
(Universidade Federal de Goiás, 2016-03-29) Maione, Camila; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Noronha, Adriana Backx; Leitão Júnior, Plínio de Sá
A practical way to characterize consumable substances is through its chemical elements in its composition and theirs concentrations. By using these elements as feature variables, it is possible to arrange these substances samples in a data matrix in which data mining and statistical techniques can be applied for predictive analysis. The classification of consumable substances based on its chemical components is an interesting problem and provides useful information for various purposes, as: recognition of geographical origin of a substance; validation and authenticity; determination of the characteristics of a product which can aid companies in the quality control and preservation; differentiation of categories of a product, and others. This study presents a methodology for predictive analysis of substances and food based on its chemical components, using data mining concepts and techniques allied to ICPMS. Four applications of the proposed methodology are described: recognition of the geographical origin of Brazilian white rice produced in SÃ£o Paulo and Goiás states; differentiation of organic and conventional Brazilian grape juice; differentiation of organic and conventional Brazilian chocolate, and analysis of its toxic and essential elements; recognition of the source of ecstasy tablets apprehended in two cities from Sao Paulo state, Ribeirão Preto and Campinas. For all applications presented, the classification models obtained showed high predictive performance (over 85%), which attest the efficiency of the proposed methodology, and the variable selection techniques used helped us to identify the chemical elements which are more important to the differentiation of the analyzed samples. For the purpose of distinguishing food samples into organic and conventional, our approach is pioneer and yielded good results.
Balanceamento de dados com base em oversampling em dados transformados
(Universidade Federal de Goiás, 2020-08-17) Maione, Camila; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Leitão Júnior, Plínio; Costa, Ronaldo Martins da; Costa, Ana Paula Cabral Seixas; Lozano, Kátia Kelvis Cassiano
Introduction: The efficiency and reliability of data analyses depends heavily on the quality of the analyzed data. The fundamental process of preparing databases in order to make them cleaner, more representative and improve their quality is called data preprocessing, during which data balancing is also performed. The importance of data balancing lies in the fact that several classification models commonly employed in enterprises and academic projects are designed to work with balanced data sets, and there are several factors which hinder classification performance which are associated to data imbalance. Objective: A new approach for data balancing based on data transformation combined with resampling of transformed data is proposed. The proposed approach transforms the original data set by transforming its input variables into new ones, therefore altering the data samples' position in the dimensional plane and consequently the choice that SMOTE-based resampling algorithms make over the initial samples, their nearest neighbours and where to place the generated synthetic samples. Methods: An initial implementation based on Principal Component Analysis (PCA) and SMOTE is presented, called PCA-SMOTE. In order to test the quality of the balancing performed by PCA-SMOTE, twelve test data sets were balanced through PCA-SMOTE and three other popular data balancing methods, and the performance of three classification models trained on these balanced sets are assessed and compared. Results: Several classification models trained on data sets which were balanced using the proposed method presented higher or similar performance measures in comparison to the same models trained on data sets that were balanced through the other evaluated algorithms, such as Borderline-SMOTE, Safe-Level-SMOTE and ADASYN. Conclusion: The satisfactory results obtained prove the potential of the proposed algorithm to improve learning of classifiers on imbalanced data sets.
Sobre conjuntos dominantes eficientes em grafos
(Universidade Federal de Goiás, 2009-03-12) Oliveira, Rommel Teodoro de; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610
Given a graph G = (V;E) and a set of vertices D V, a vertice v 2 V is dominated by D if jN[v] \ Dj 1. When jN(v) \ Dj = 1 for all v 2 V, G is efficiently dominable. A generalization of this concept is called efficient multiple domination, which requires all vertices must be dominated by a set D V exactly k times. The aim of this dissertation is to study these topics, describing the theoretical knowledge needed for advanced researches. For this reason, many of the theorems and its proofs are detailed. Furthermore, some results on the efficient multiple domination are presented, including bounds for the size of efficient k-dominating sets, the complement and iterated line graphs of efficiently (r + 1)-dominable r-regular graphs and a N P-completeness proof for the efficient multiple domination problem in arbitrary graphs. It is expected that this work contribute to the development of future researches on the efficient domination and in the resolution of some open problems.
Abordagem de seleção de características baseada em AUC com estimativa de probabilidade combinada a técnica de suavização de La Place
(Universidade Federal de Goiás, 2023-09-28) Ribeiro, Guilherme Alberto Sousa; Costa, Nattane Luíza da; http://lattes.cnpq.br/9968129748669015; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Lima, Marcio Dias de; Oliveira, Alexandre César Muniz de; Gonçalves, Christiane; Rodrigues, Diego de Castro
The high dimensionality of many datasets has led to the need for dimensionality reduction algorithms that increase performance, reduce computational effort and simplify data processing in applications focused on machine learning or pattern recognition. Due to the need and importance of reduced data, this paper proposes an investigation of feature selection methods, focusing on methods that use AUC (Area Under the ROC curve). Trends in the use of feature selection methods in general and for methods using AUC as an estimator, applied to microarray data, were evaluated. A new feature selection algorithm, the AUC-based feature selection method with probability estimation and the La PLace smoothing method (AUC-EPS), was then developed. The proposed method calculates the AUC considering all possible values of each feature associated with estimation probability and the LaPlace smoothing method. Experiments were conducted to compare the proposed technique with the FAST (Feature Assessment by Sliding Thresholds) and ARCO (AUC and Rank Correlation coefficient Optimization) algorithms. Eight datasets related to gene expression in microarrays were used, all of which were used for the cross-validation experiment and four for the bootstrap experiment. The results showed that the proposed method helped improve the performance of some classifiers and in most cases with a completely different set of features than the other techniques, with some of these features identified by AUC-EPS being critical for disease identification. The work concluded that the proposed method, called AUC-EPS, selects features different from the algorithms FAST and ARCO that help to improve the performance of some classifiers and identify features that are crucial for discriminating cancer.
Abordagem de seleção de características baseada em AUC com estimativa de probabilidade combinada a técnica de suavização de La Place
(Universidade Federal de Goiás, 2024-09-28) Ribeiro, Guilherme Alberto Sousa; Costa, Nattane Luíza da; http://lattes.cnpq.br/9968129748669015; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Lima, Marcio Dias de; Oliveira, Alexandre César Muniz de; Gonçalves, Christiane; Rodrigues, Diego de Castro
The high dimensionality of many datasets has led to the need for dimensionality reduction algorithms that increase performance, reduce computational effort and simplify data processing in applications focused on machine learning or pattern recognition. Due to the need and importance of reduced data, this paper proposes an investigation of feature selection methods, focusing on methods that use AUC (Area Under the ROC curve). Trends in the use of feature selection methods in general and for methods using AUC as an estimator, applied to microarray data, were evaluated. A new feature selection algorithm, the AUC-based feature selection method with probability estimation and the La PLace smoothing method (AUC-EPS), was then developed. The proposed method calculates the AUC considering all possible values of each feature associated with estimation probability and the La Place smoothing method. Experiments were conducted to compare the proposed technique with the FAST (Feature Assessment by Sliding Thresholds) and ARCO (AUC and Rank Correlation coefficient Optimization) algorithms. Eight datasets related to gene expression in microarrays were used, all of which were used for the crossvalidation experiment and four for the bootstrap experiment. The results showed that the proposed method helped improve the performance of some classifiers and in most cases with a completely different set of features than the other techniques, with some of these features identified by AUC-EPS being critical for disease identification. The work concluded that the proposed method, called AUC-EPS, selects features different from the algorithms FAST and ARCO that help to improve the performance of some classifiers and identify features that are crucial for discriminating cancer.
Aprimoramento do modelo de seleção dos padrões associativos: uma abordagem de mineração de dados
(Universidade Federal de Goiás, 2021-12-20) Rodrigues, Diego de Castro; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Costa, Ronaldo Martins da; Costa, Nattane Luíza da; Rocha, Marcelo Lisboa; Jorge, Lúcio de Castro
The objective of this study is to improve the association rule selection model through a set of asymmetric probabilistic metrics. We present the Health Association Rules - HAR, based on Apriori, the algorithm is composed of six functions and uses alternative metrics to the Support/Confidence model to identify the implication X → Y . Initially, the application of our solution was focused only on health data, but we realized that asymmetrical associative patterns could be applied in other contexts that seek to address the cause and effect of a pattern. Our experiments were composed of 60 real datasets taken from specialist websites, research partnerships and open data. We empirically observed the behavior of HAR in all data sets, and a comparison was performed with the classical Apriori algorithm. We realized that it has overcome the main problems of the Support/Confidence model. We were able to identify the most relevant patterns for the observed datasets, eliminating logical contradictions and redundancies. We also perform a statistical analysis of the experiments where the statistical effect is positive for HAR. HAR was able to discover more representative patterns and rare patterns, in addition to being able to perform rule grouping, filtering and ranking. Our solution presented a linear behavior in the experiments, being able to be applied in health, social, content suggestion, product indication and educational data. Not limited to these data domains, HAR is prepared to receive large amounts of data by using a customized parallel architecture.
Sobre alianças defensivas e ofensivas globais em alguns produtos de grafos e grafos simpliciais
(Universidade Federal de Goiás, 2015-10-30) Silva, Leila Roling Scariot da; Dourado, Mitre Costa; http://lattes.cnpq.br/0841425239502177; Barbosa, Rommel Melgaço; http://lattes.cnpq.br/6228227125338610; Barbosa, Rommel Melgaço; Dourado, Mitre Costa; Federson, Fernando Marques; Rosa, Thierson Couto; Santos, José Plínio de Oliveira
Given a graph G, a defensive alliance of a set of vertices A⊆V(G) satisfying the condition that for each v ∈ A, |N[v] ∩ A| ≤ |N[v] − A|. The set S is an offensive alliance if the inaquality holds for every v ∈ N[S]−S. A alliance A is called global if is also a dominant set. In this paper, we establish lower bounds for Simplicial Graphs and further give closed formulas and upper bounds to decide the global, defensive, offensive, alliance numbers for lexicographic product of paths, cycles, stars and complete graphs. We establish a relationship to global defensive alliance numbers and complementary prism product to graphs.

Navegar

Navegando INF - Instituto de Informática por Por Orientador "Barbosa, Rommel Melgaço"

Resultados por página

Opções de Ordenação