Call for IPASJ Journals Coming Issues

Welcome to IPASJ!

IIJCS:Volume 6, Issue 4, April 2018

_____________________________________________________________________________

Title:
IMPROVED ENSEMBLE SUBSPACE CLASSIFIER WITH HIGHER PRECEDENCE ACTIVE SAMPLING ON CLASS ASSOCIATION RULES
Author Name:
Kayal Padmanandam, Kannan Subramaniam
Abstract:
ABSTRACT Ensemble classifier an assemblage of classifiers, is a prospective practice for improvising the capacity of individual classifiers. Ensemble classifier has produced a powerful accuracy related to single classifiers. The aim of this work is to explain the role of sampling done on class association rule to enhance the predictive power of the algorithm Ensemble Classification on Fuzzy Utility Mining with active sampling(ECFUM-AS). This classifier learning system uses random subspace algorithm that outfits training of increased attributes, despite many fuzzy classifications escalates attribute magnification. The sampling technique used here for accuracy improvisation can play a better prediction in the decision-making system. As soon as the research essences for better accuracy factor, we focused on subspace sampling. Subsequently, the infinite sampling space is created only with class association rules with higher order confidence precedence, which has agreed an immense progress in algorithm’s accuracy. Results elevated the accuracy about 6 to 7 percent on increased rule set. Keywords: Association Rules, Classification algorithm, Sampling Methods, Knowledge based System, Knowledge Discovery
Cite this article:
Kayal Padmanandam, Kannan Subramaniam , " IMPROVED ENSEMBLE SUBSPACE CLASSIFIER WITH HIGHER PRECEDENCE ACTIVE SAMPLING ON CLASS ASSOCIATION RULES" , IPASJ INTERNATIONAL JOURNAL OF COMPUTER SCIENCE(IIJCS) , Volume 6, Issue 4, April 2018 , pp. 007-012 , ISSN 2321-5992.
Full Text [PDF]                                                                                                          Back to Current Issue

________________________________________________________________

International Publisher for Advanced Scientific Journals (IPASJ)
email Id: publisher@ipasj.org

 

 

©  2010 All Rights Reserved to IPASJ