Sandro Rautenberg

Fuzzy expert system for predicting pathological stage of prostate cancer

Prostate cancer is the second most common cancer among men, responsible for the loss of half a million lives each year worldwide, according to the World Health Organization. In prostate cancer, definitive therapy such as radical prostatectomy, is more effective when the cancer is organ-confined. The aim of this study is to investigate the performance of some fuzzy expert systems in the classification of patients with confined or non-confined cancer. To deal with the intrinsic uncertainty about the variables utilized to predict cancer stage, the developed approach is based on Fuzzy Set Theory. A fuzzy expert system was developed with the fuzzy rules and membership functions tuned by a genetic algorithm. As a result, the utilized approach reached better precision taking into account some correlated studies.

DBtrends: Exploring Query Logs for Ranking RDF Data

Many ranking methods have been proposed for RDF data. These methods often use the structure behind the data to measure its importance. Recently, some of these methods have started to explore information from other sources such as the Wikipedia page graph for better ranking RDF data. In this work, we propose DBtrends, a ranking function based on query logs. We extensively evaluate the application of different ranking functions for entities, classes, and properties across two different countries as well as their combination. Thereafter, we propose MIXED-RANK, a ranking function that combines DBtrends with the best-evaluated entity ranking function. We show that: (i) MIXED-RANK outperforms state-of-the-art entity ranking functions, and; (ii) query logs can be used to improve RDF ranking functions.

LODFlow: a workflow management system for linked data processing

The extraction and maintenance of Linked Data datasets is a cumbersome, time-consuming and resource-intensive activity. The cost for producing Linked Data can be reduced by a workflow management system, which describes plans to systematically support the lifecycle of RDF datasets. We present the LODFlow Linked Data Workflow Management System, which provides an environment for planning, executing, reusing, and documenting Linked Data workflows. The LODFlow approach is based on a comprehensive knowledge model for describing the workflows and a workflow execution engine supporting systematic workflow execution, reporting, and exception handling. The environment was evaluated in a large-scale real-world use case. As result, LODFlow supports Linked Data engineers to systematically plan, execute and assess Linked Data production and maintenance workflows, thus improving efficiency, ease-of-use, reproducibility, reuseability and provenance. The environment was evaluated in a large-scale real-world use case. As result, LODFlow supports Linked Data engineers to systematically plan, execute and assess Linked Data production and maintenance workflows, thus improving efficiency, ease-of-use, reproducibility, reuseability and provenance.

Apoiando estudos cientométricos com Linked Open Data

Na Cientometria, mensurar indicadores e uma tarefa complexa devido aos desafios em coletar, organizar e relacionar dados. Tais desafios permeiam a web, onde os dados sao distribuidos em varias fontes e formatos incompativeis. Esses problemas podem ser resolvidos com o emprego de tecnologias e metodologias baseadas nos principios Linked Open Data. Tais principios sao fundamentados num conjunto de melhores praticas de Web Semântica e Dados Abertos para organizar, publicar e conectar dados na web. Atraves destes, dados sao acessados e consumidos sem restricoes, em diversas aplicacoes. No presente trabalho, aborda-se a disponibilizacao do historico do indice Qualis conforme o Linked Open Data. Pressupoe-se que tal empreendimento e importante na reutilizacao de dados em pesquisas bibliometricas/cientometricas, servindo para: mensurar a evolucao dos periodicos cientificos; auxiliar na afericao de medidas qualitativas e quantitativas de publicacoes cientificas; ou obter informacoes relevantes a partir do cruzamento com outros indicadores cientometricos. A disponibilizacao do indice Qualis e verificada em tres estudos de caso. Como resultados, tem-se a disponibilizacao aberta do historico do indice Qualis (dos anos 2005 a 2013) e de interfaces web para acesso e consumo dos dados.

Natural Computing in Pump-Scheduling Optimization for Water Supply System: Case Study

The electrical energy cost represents a significant fraction of the total cost in a water supply system. Any optimization in pumping operational procedures results in a reduction of this cost. The aim of this paper is the optimization of pump operation in a water distribution system, located at Guarapuava, Brazil. For this, we used two techniques of Natural Computing: Genetic Algorithms and Shuffled Frog Leaping Algorithm. Both techniques were effective when comparing with a traditional approach. However, in our experiments, the SFLA achieved lower costs.

Training strategies for Radial Basis Function Neural Networks: A study applied to prostate cancer prognosis

A development of a Radial Basis Function Neural Network applied to the prostate cancer prognosis is presented. Five training strategies were implemented: the Successive Approximation algorithm; the k-means algorithm; Ainet; Ainet + k-means, and Successive Approximation + k-means. Comparing all tested strategies, the Successive Approximation training strategy obtained the best fitness measure. And comparing this result to other previously studies, we concluded that the use of Radial Basis Function Neural Networks becomes a viable alternative to the prostate cancer prognosis.