Dimensionality Reduction: An Effective Technique for Feature Selection

Abstract

For knowledge gaining the dimensionality reduction is a significant technique. It has been observed that most of the time dataset is multidimensional and larger in size. When we are using same dataset for classification it may create wrong results and it may also requires more requirements in terms of storage as well as processing capability. Most of the features present are redundant, inconsistent and degrade the performance. To increase the effectiveness of classification these duplicate and inconsistent features must be removed. In this research we have introduced a new method for dealing with the problem of dimensionality reduction. By reducing the unrelated (irrelevant) and unnecessary features related to data, or by means of effectively merging original features to produce a smaller set of feature with more discriminative control, dimensionality reduction methods convey the instant effects of rapid the data mining algorithms, better performance, and increase in unambiguous of data model