Kannada Alphabets Recognition using Decision Tree and Random Forest Models

Character recognition system is the most emerging technology for digitizing handwritten documents from any type of script. The classification has a variety of approaches and the techniques of classification are preferred as they utilize a combination of different methods of classification to give a more accurate machine-learning model. One such ensemble classification technique is Random Forest, which aims to classify data with the help of several decision trees. This article, describes the whole process of alphabet recognition from preprocessing of the datasets in creating a model for binary and multiclass classification using MATLAB and is aimed at highlighting the advantages of random forest classification on a single decision tree for prediction of handwritten letters.