A review of dimensionality reduction methods applied on clinical data of diabetic neuropathy complaints

Abstract

Aim of this paper is to apply Dimensionality Reduction techniques on a large clinical data of Type II diabetes patients in identifying the causes and symptoms that they tend to develop neuropathic complications. Data preprocessing is an essential technique using Machine Learning (ML) and Data Mining are ineffective for big data analytics. An effective approach to dimensionality reduction is reduction of the number of independent/dependent variables which are necessary for our analysis. These processes help in identifying the features to be considered for selection and extraction avoiding the redundant and irrelevant features by choosing subset features which are a linear combination of the original. Both supervised and unsupervised learning are applied for prediction and further analysis. This paper primarily focuses on Supervised Learning, the variables are known beforehand. Combination of feature selection techniques and ML algorithms are used to support practitioners with the best methods for feature reduction and extraction.