Parallel Distributed Implementation of Multiobjective Genetics-based Machine Learning Algorithms

2015 Fiscal Year Final Research Report

• Project/Area Number: 25330292
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Soft computing
• Research Institution: Osaka Prefecture University
• Principal Investigator: Nojima Yusuke 大阪府立大学, 工学(系)研究科(研究院), 准教授 (10382235)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: 知識獲得 / 多目的最適化 / 並列分散実装 / 進化計算

Outline of Final Research Achievements

There are two goals for data mining from numerical data. One is to maximize the accuracy of obtained knowledge. The other is to maximize its interpretability. However, there is a tradeoff between the accuracy and the interpretability. To address this issue, we proposed multiobjective genetics-based machine learning (MoGBML) which can simultaneously handle these two objectives and provide a number of classifiers with different accuracy and interpretability as knowledge. To further extend MoGBML to large data sets, we apply our parallel distributed implementation to MoGBML in this study. In addition, we examine the effects of a various kind of antecedent sets on the performance of our parallel distributed GBML. We also examine the applicability of our parallel distributed implementation to data mining from multiple data centers.

Free Research Field

計算知能