Performance Analysis of KNN and SVM Classifiers Using Handwritten Kannada Vowels Recognition

This work emphasises on the development of Kannada vowels character recognition system using KNN and SVM and performs a recognition performance analysis for both models. The main goal of this project is mainly to compare the performance of two classifiers i. e. KNN (k- nearest neighbor) and SVM (Support Vector Machine) and to obtain their performance plot. A GUI which is integrated with the binaries of KNN/LIBSVM and language rules (stores the set of valid strokes which makes a character) are used, testing is done. The classifiers performance is measured as classification accuracy like correct rate and error rate. Initially the classifiers are being trained with the training samples obtained from various users then the classifiers are tested on a test samples obtained from the users and the performance is being noted and plotted, by observing this plot can tell which classifier performance is more and better suited for the recognition application and the documented text will be converted into machine editable format. Here KNN outperforms well than the SVM. In this method the GUI is developed to show the overall recognition rates and plots. KNN gives 100 % accuracy where SVM gives only 92.56 % accuracy.