SUNRISE: Um Algoritmo Rápido de Aprendizado Multiestratégia

Aloísio Carlos de PinaGerson Zaverucha

This work presents the analysis, implementation and experimental evaluation of a new fast multi-strategy learning algorithm. The algorithm was compared to its predecessor by using an approximate statistical test, the n-fold cross-validated paired t-test. Many data sets were used in the tests, in an attempt to include several domains, sizes and difficulties. Comparing the average execution times of the two algorithms, it could be verified that the new algorithm is faster and possesses similar accuracy to the one of its predecessor, confirming that the objective of the research was reached. Clique no link abaixo para buscar o texto completo deste trabalho na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: