• Ms. Chapala Maharana
• Ch. Sanjeev Kumar Dash
• Bijan Bihari Mishra

It has been realized that feature subset selection is running with a lot of challenges even though many efforts have been made so far for small to large scale dimensional datasets. But for larger dimensional datasets a very rare feature subset selection techniques have been developed with their own set of constraints. In this work, we propose a novel two steps feature subset selection for high dimensional datasets. In the first step the features in the database are ranked and are arranged as per their rank from maximum to minimum. Then few maximum ranking features are presented to Extreme Learning Machine (ELM) for classification and the minimum subset that performs well is selected. The proposed approach is evaluated with a few benchmarking highly skewed dataset retrieved from University of California, Irvine (UCI) repository. In addition to classification accuracy, four other performance metrics such as sensitivity, specificity, Jaccard index and M-Estimate also used to validate the results. The experimental study is encouraging us to pursue further research in high dimensional skewed data.

Extreme Learning Machine; signal to noise ration; classification; feature selection

"Feature Selection using ELM Classifier", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.7, Issue 6, page no.769-774, June-2020, Available :http://www.jetir.org/papers/JETIR2006103.pdf

"Feature Selection using ELM Classifier", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.7, Issue 6, page no. pp769-774, June-2020, Available at : http://www.jetir.org/papers/JETIR2006103.pdf