Exploring the effect of normalization on medical data classification
Namrata Singh, Pradeep Singh
Abstract
Data normalization as one of the pre-processing strategies is utilized either to transform or scale the data in order to make an equal contribution of each attribute. For a given classification problem, the performance of any machine learning approach depends upon the quality of data in order to produce a generalized classification approach. Various studies have shown the significance of data normalization to enhance the quality of data and finally the performance of machine learning techniques. But there is dearth of investigations about the effect of data normalization methods in classifying the medical datasets. Thus, this study intends to explore the effect of three data normalization techniques namely min-max, z-score and Median and Median Absolute Deviation on the performance of four classification algorithms namely Naïve Bayes, Support Vector Machine - Radial Basis Function, Random Forest and k-Nearest Neighbour. The experiments conducted on 20 publicly available medical datasets are based on the classification accuracy as performance parameter. The best performance results were obtained with z-score normalization method along with Random Forest classifier.