A Feature Selection process Optimization in multi-class Miner for Stream Data Classification
Journal: INTERNATIONAL JOURNAL OF COMPUTERS & TECHNOLOGY (Vol.3, No. 3)Publication Date: 2013-01-01
Authors : Manish Rai; Rekha Pandit; Vineet Richhariya;
Page : 359-364
Keywords : Stream Data classification; MCM; AGA and MGM-GA;
Abstract
Multi-class miner resolves the problem of feature evaluation, data drift and concept evaluation of stream data classification. The process of stream data classification in multi-class miner based on ensemble technique of clustering and classification on feature evaluation technique. The process of feature evaluation technique faced a problem of correct point selection of cluster centre for the process of data grouping. For the proper selection of features point we used optimization technique for feature selection process. The feature selection process based on advance genetic algorithm (AGA). The advance genetic algorithm poses a process of feature point for neighbour class detection for finding a correct point in classification. Our proposed algorithm tested on some well know data set provided by UCI machine learning repository. Our empirical evaluation result shows that better result in comparison of multi-class miner for stream data classification.
Other Latest Articles
- ANALYSIS OF STUDENT ACTIVITIES ON COMPUTER - A STUDY ON PANJAB UNIVERSITY, INDIA
- Performance Evolution of Intrusion Detection system on MANET Using Genetic Evolution
- Influence of Biofield Treatment on Physical, Structural and Spectral Properties of Boron Nitride
- An Impact of Biofield Treatment on Spectroscopic Characterization of Pharmaceutical Compounds
- Effect of Biofield Treatment on Structural and Morphological Properties of Silicon Carbide
Last modified: 2016-06-30 14:13:24