A Review on Dimension Reduction Techniques in Data Mining

Abstract

Real world data is high-dimensional like images, speech signals containing multiple dimensions to represent data. Higher dimensional data are more complex for detecting and exploiting the relationships among terms. Dimensionality reduction is a technique used for reducing complexity for analyzing high dimensional data. There are many methodologies that are being used to find the Critical Dimensions for a dataset that significantly reduces the number of dimensions. They reduce the dimensions from the original input data. Dimensionality reduction methods can be of two types as feature extractions and feature selection techniques. Feature Extraction is a distinct form of Dimensionality Reduction to extract some important feature from input dataset. Two different approaches available for dimensionality reduction are supervised approach and unsupervised approach. One exclusive purpose of this survey is to provide an adequate comprehension of the different dimensionality reduction techniques that exist currently and also to introduce the applicability of any one of the prescribed methods that depends upon the given set of parameters and varying conditions. This paper surveys the schemes that are majorly used for Dimensionality Reduction mainly high dimension datasets. A comparative analysis of surveyed methodologies is also done, based on which, best methodology for a certain type of dataset can be chosen.

Keywords: Data Mining, Dimensionality Reduction, Clustering, feature selection; curse of dimensionality; critical dimension