Multithreaded and Spark parallelization of feature selection filters

详细信息    查看全文

• 作者：Carlos Eiras-Franco^a ; ^{carlos.eiras.franco@udc.es}Author Vitae ; Veró ; nica Boló ; n-Canedo^aAuthor Vitae ; Sabela Ramos^bAuthor Vitae ; Jorge Gonzá ; lez-Domí ; nguez^bAuthor Vitae ; Amparo Alonso-Betanzos^aAuthor Vitae ; Juan Touriñ ; o^bAuthor Vitae
• 关键词：Multithreading ; Spark ; Feature selection ; Machine learning
• 刊名：Journal of Computational Science
• 出版年：2016
• 出版时间：November 2016
• 年：2016
• 卷：17
• 期：part_P3
• 页码：609-619
• 全文大小：2419 K
• 卷排序：17

文摘

Multithreaded and Spark implementations for feature selection algorithms proposed. Algorithms included are RELIEF-F, InfoGain, CFS and SVM-RFE. Superlinear performance gains obtained for RELIEF-F. Spark implementation for CFS much more efficient than the one included in Weka.