Improvised number identification using SVM and random forest classifiers

Abstract

The classification of any dataset is the common task of any algorithm. In machine learning, data pre-processing is essential steps for quality of data and information. The ability of classifier to learn directly depends on quality of data. Therefore, in some case, it is very important that the data need to pre-process before feeding it to the classifier. Normalization and Standardization are the pre-processing methods which we have used. The paper deals with the methods for improvised number identification analyse the pre-processing methods on Support Vector Machine (SVM) and Random Forest (RF) on the handwritten digit dataset. It is important to identify appropriate pre-processing method for specific data sets and algorithms. As a result, it is shown that using pre-processing we can improve the accuracy of SVM for hand written digit classification.

Subject Classification:

• Primary 62J05
• Secondary 03C45

Keywords:

• Support Vector Machine
• Random Forest
• Normalization
• Standardization