DIFFICULTY IN HANDLING HIGH DIMENSIONAL DATA- PLEASE STAND UP

Shikha Agarwal, R. Rajesh

Abstract: With the advancement of technology huge amount of high dimensional data are getting generated. This paper present’s the high dimensional data as a serious problem for computational analysis. Broadly high dimensional data could be handled in two ways, firstly with dimensionality reduction, secondly without dimensionality reduction using classical machine learning. Two main categories of dimensionality reduction are feature selection and extraction. Feature selection and feature extraction. Feature selection selects a subset of features based on some criteria while feature extraction method transforms the data into the lower dimensional space. Experiment result of the high dimensional data shows the performance improvement of classifier before and after dimensionality reduction.

Keywords: Feature Selection, Feature Extraction, Wrapper Method, Embedded Methods, IBPSO

DOI: https://doi.org/10.15623/ijret.2014.0326022