Repositório RCAAP

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

Fundação de Amparo a Pesquisa do Estado de Minas Gerais

Conselho Nacional de Desenvolvimento Científico e Tecnológico

The principal objective of an image retrieval system is to obtain images which are as similar as possible to the user´s requirements, from all the images in the reference collection. Such an objective is difficult to reach due principally to the subjectivity of the image similarities. This is due to the fact the images can be interpreted in different ways by different people. With the aim of resolving this problem the content-based image retrieval systems explore the features of color, shape and texture. These are nearly always associated to the regions and use relevance feedback mechanisms to adjust a search to the user s criterions. Various approaches have been used in relevance feedback from those genetic algorithms have become quite popular due to their adaptive abilities. In this work we presented an image retrieval system based on the similarity of local patterns, working with the features of color, shape and texture as well as relevance feedback via a genetic algorithm. The task of this algorithm is infer weights to the features of color, shape, texture and regions which better adjust to the similarity found between images through the user s search criterions, thus producing a final ranking which is in accordance with the criterions expressed in the relevance feedback. The genetic algorithms theory states that the fitness measure applies an essential role upon the performance of these algorithms, once the fitness measure directs the search path for the evaluation of each individuals aptitude. Due to the lack of consensus about the best fitness measure in the ranking evaluation problem we present a performance analysis of ten fitness functions. The fitness functions are classified in two groups: order-based and non-order based. Some of these functions are adapted from textbased information retrieval systems and others are proposed in this work. The experimental results show that the order based fitness functions are more compatible to the user s interests, once they present superior rankings in terms of precision for low recall rates and conduct the quickest genetic algorithm in the search for an optimal heuristic solution. The results obtained are superior to those of the works of Stejic et al., which served as our inspiration.

Artificial Intelligence techniques are increasing their role in the solution of biological problems. The present study use a Genetic Algorithm (GA) in the search for high level IF-THEN rules. This GA was applied to miner classification rules from a gene expression database named NCI60. This database was developed using cancer cells measured by microarray. The goal of this mining is the discovery of relations among gene expression level and the nine types of cancer classes analyzed in this work.

The objective of this work is to propose an approach based in hybrid systems for study of biological process. Initially a study about mathematical modelling of systems was carried out looking for appropriate tools for modelling biological process. Various models of biological process was found in literature using different approaches (differential equations, hybrid automat, and Petri nets). The hybrid model used in this work is the differential predicate transition Petri nets. The differential predicate transition Petri nets combine continuous and discrete aspects are represented by a set of differential equations and the hybrid aspects are represented by a predicate transition Petri net. Based in this approach, three study cases was choose to to show the validate this approach. This three study cases was modelled using the differential predicate transition Petri nets and these models was simulated by using MatLab. The results obtained with the simulations provides information for a quantitative analysis of the models, validating the differential predicate transition Petri nets for modelling biological process.

Discovering frequent patterns in databases is an important problem for knowledge discovery and its importance is justified by the diversity of areas where it can be used, such as retail, financial market, medicine, agriculture, farming, telecommunications, etc. The problem of mining frequent patterns in temporal databases, also know as mining temporal patterns, has been widely studied. In some works the temporal patterns are expressed by propositional temporal logic, and in others, the temporal patterns are more expressive and are expressed by first order temporal logic. The existing temporal patterns in the literature represent the time either in terms of points, where event occur in determined instants, or in terms of intervals, where event occur during a period of time. This dichotomy implies that some facts may not be inferred. For instance, facts related to the clinical history of a patient, where, the patient took some medicine during a period of time and was submitted to a surgery on a day, during the period when he or she was taking the medicine. In this dissertation, we propose a new temporal pattern, called the hybrid temporal pattern or simply htp, where, time is represented in terms of points and/or intervals. This pattern is very expressive. We adapted Allen's Interval Temporal Logic to specify it. We also proposed the algorithm MILPRIT* for mining the frequents htp's in a database with respect to a minimum support and satisfying a constraint specified by user through a regular expression. Doing so, we allow the user to control the process of htp discovery. The performance and scalability of MILPRIT* has been evaluated through a set of experiments over synthetic and real databases.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

Most work on pattern mining focus on simple data structures like itemsets or sequences of itemsets. However, a lot of recent applications dealing with complex data like chemical compounds, protein structure, social network, XML and Web Log databases, require much more sophisticated data structures (trees or graphs) for their specification. Here, interesting patterns involve not only frequent object values (labels) appearing in the trees (or graphs) but also frequent specific topologies found in these structures. Mining frequent tree patterns have been extensively studied, motivated by the increasing interest and applicability in different areas (Web Mining, Bioinformatics, etc). However, conventional tree mining systems normally consider only minimum support criterium as a mechanism for filtering patterns to be mined. After mining process, hard work is requiring to filter patterns concerned with user interests. In this dissertation, we propose CobMiner, Constrained-based Miner, a tree pattern mining algorithm which incorporates tree automata into the mining process in order to restrict the mining scope and to generate frequent patterns more closely related to user interests. We compare two methods for introducing user constraints into the discovery process: the first one is CobMiner which incorporates tree automata constraints as an intra-mining mechanism, the second one is TreeMinerPP which consists of a well-known tree pattern mining algorithm, TreeMiner, followed by a post-processing phase, where patterns are filtered using a tree automatum. An extensive set of experiments executed over synthetic and real data (XML documents) allow us to conclude that incorporating constraints during the mining process is far better effective than filtering the frequent and interesting patterns after the mining process.

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior

This work presents an overview of INDIAM INterpretation and DIAgnosis of Mammograms an e-learning system for assisting medical students in the interpretation of mammograms and diagnosis of breast cancer. The knowledge base of INDIAM is composed of a mammographic database that includes cases with all the available mammographic standard views, global and local radiological findings, diagnosis proven by biopsy, the patient s clinical history, and information about the life style of the patient; an ontology, called BreastCancerOnto, that available semantics in the Web services integrated into the e-learning system; and a hypertext tutorial which guides the users (doctors, students, or researchers) through the concepts related to the diagnosis of breast cancer. The proposed e-learning system is composed of four principal Web services: one to guide the beginner student through the steps of analyzing a mammogram; one to simulate the diagnosis of a given mammogram retrieved from the mammographic database; one to answer questions from the user based on BreastCancerOnto ontology; and one to make available the tutorial about diagnosis of breast cancer. The INDIAM e-learning system is being integrated into AMDI - an indexed atlas of digital mammography that can be accessed via the Web.

Computing is causing a revolution in peoples and business life. Internet is for sure the reason why this ubiquitous computing is present in all almost all daily process. Web Services offers a new way to use Internet, where remote applications can communicate each other using interoperable standards. Through this approach the Web can be the source not of only information, but distributed computing. In this environment security is a critical issue and the security model to Web Services is not fully defined. Access control is a way to guarantee that only the users, or systems, that are fully recognized can access services and information. Passwords are a spread way used to accomplish this. Nevertheless, most systems currently used to perform this control shows deficiencies and an intruder can break it down. This work presents and deploys a solution for the Web Services authentication control. This solution is based on SRP (Secure Remote Password), a strong-based password authentication protocol. To accomplish this, the protocol has to be extended in offer authentication control as an independent layer can be built on any available Web Services plataform. In this work we have used Apache AXIS for this end.

This dissertation walks through the Advantage Function Model, from Goldwasser and Bellare, explaining its complex concepts and exemplifying its application in the analysis of encryption security schemes currently used in practice. It is also shown ways to put this model into practice. The safety of the schemes used in the IEEE 802.11 Standard are explained and used as the object of study in the model application. The results can be used to justify the adoption of certain security technologies. The analysis presented demonstrates the great value of its study and of the use of mathematical models in safety demonstrations. It can help designers and developers to maximize the safety of their encryption schemes, by the effective application of the model or by the simple understanding of the concepts of security mentioned here.

This dissertation proposes a password presentation system in hostile machine, based on Coding Theory or, more specifically, the standard array concept. The objective is to permit the user to inform his password to a server with no key in or to inform in a explicit manner. The system aim is to guarantee the user s security keeping his password protected in the whole communication process, considering the fact that the user s machine is hostile.

This work presents a new approach that uses Artificial Intelligence Planning techniques in the development of a Workflow Modelling System. This approach concentrates on the field of process modelling. The planning tool is used to automatically generate models of consistent processes. The modeled processes usually have activities that split the flow into more than one possible branch. The use of planner to perform this modelling process, exacts conditional plans. This work presents algorithms give rise to conditional plans when using classic planners. Finally, experiments are carried out to measure the efficiency and effectiveness of these algorithms, and the results are compared to those of a contingent state-of-the-art planner.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais

Image retrieval is an important research area in Digital Image Processing and Computa tional Vision that can be applied in many areas such as medical images diagnosis, crime prevention, personal identification (finger-print), intelectual property, etc. The content based image retrieval systems (CBIR) has as the main goal of retrieving images in image database that are more similar with a query image chosen by the user, based on the features automatically extracted from the images. The appearance of the CBIR systems can be justified by the fact that traditional indexation methods based on text, require much more time and efforts in the indexation process for huge images databases. The most explored visual features in CBIR are color, texture and shape. Concerning to texture, there are three main approaches: a statistical, a structural and a spectral one. The statistical approach considers the color distribution and their inter-relationship. The structural approach, by the other side, works with spatial arrange of structural primitives, while the spectral approach is based on the spectral frequency properties, obtained through transformations such as Fourier and Wavelets. Following the ideas presented by Shapiro and Brady and Carcassoni, Ribeiro and Hancock, this work explores how the modal structure of the pattern, taken in the textures' frequency space can be used for retrieval purposes. Carcassoni, Ribeiro and Hancock presented a variation of the correspondence method of Shapiro and Brady, that aims to match point sets by comparing the eigenvectors of a matrix that measures the inter-relationship between the pairwise points (proximity ma- trix). Carcassoni, Ribeiro and Hancock introduces a texture descriptor based on the image power spectrum peaks, with the aim of represent it. In this work, was used a variation of the Lloyd's quantization technique from the image power spectrum to represent it. With the aim of verifying the method e±ciency, several experiments were carried out using regular, non-regular, homogeneous and non-homogeneous textures. The image collection is composed of images such as tissue, fabric paper, landscapes, vehicles, wooden °oor, bricks, buildings images, etc, that were extracted from several images database. The results obtained by the proposed method are compared with the Carcassoni's method and also with the gray level co-occurrence matrix method of Haralick, that is a well-known and a method widely used for texture feature extraction. The performance of the three compared methods is measured by the commonly used retrieval performance measurement, precision and recall, which is considered one of the most important techniques for performance analysis of any retrieval systems.

Fundação de Amparo a Pesquisa do Estado de Minas Gerais

The General Objective of this study is investigate and analyze the possible points that highlight the adherence of the Interorganizational Cost Management (ICM), in the production chain of sugar cane, for the management of the total costs in a sugarcane mill in the Triangulo Mineiro, hereby named Destak Mill. The research was characterized as follows: concerning its nature, how applied; concerning its objective, how exploited; concerning procedures, as a study of a unique case; and, concerning the problem, as qualitative. For the study of this case a method of semi-structured research and interviews was applied, after which transcription of the same, and analysis of contents. For validation of this case study, a process of triangulation of methods for bibliographic survey was used, besides documental research and unsystematic observation. After having carried out these methods the result presented demonstrated a non-formal existence in a ICM at Destak Mill, but in an informal restricted manner showing some factors with made impossible its effective application as a Strategic Cost Management (SCM) in management of total costs. In conformity with the categories analyzed, certain indications concerning adherence to the ICM in the process of management of total costs, observed in relationship variables (interdependence, co-operation, trust and stability/mutual benefits), however there was a lack of sharing cost information in a strategic manner. Therefore, since we are dealing with a mill that is still being developed, once it reaches its maximum productive capacity we must point out, for future studies, a possible change in results in mills at different life cycles.

The Monday effect is a calendar anomaly of the interest of researchers around the world, who try to identify behavioral or rational reasons to understand the emergence of this recurrence standard. It is considered that this effect is legitimized in the stock market; however, a provocation arises to discover if the persistence (or not) of this effect is observed in a continued growth industry in Brazil, in the stock investments funds. Thus, this research aims to analyze the relationship of Monday with the return on investments funds. Therefore, the sample was composed by 174 Brazilian investments funds totaling 134.584 observations for daily data in the period from 1st January 2011 to 11 July 2014. Using regression analysis with stock data and considering control variables identified as relevant in literature, the main results showed that the Monday effect also occurs in the investment funds segment. Furthermore, the following notes are highlighted: (i) evidences that managers can obtain abnormal returns by exploiting the Monday effect in investments funds; (ii) indications that the fund manager in the decision-making process is influenced by behavioral biases and mental shortcuts, presenting tendentious behaviors and not following a rationality; (iii) managers can try to get extraordinary returns through negotiations conducted with information from the days of the week; and (iv) IBOVESPA variable had a positive sign, i.e., the funds returns of the sample tend to move in the same direction of IBOVESPA returns. This aspect indicates that the sample equity funds has significant investments in companies with higher liquidity indexes. For future studies researches on other types of investment funds are recommended, identifying the impact of taxation on investment funds and the application of questionnaires and/or interviews to seek empirical findings in which the behavior of the manager can influence the emergence and the persistence of anomalies.