Multi-Layer Perceptron neural network and Support Vector Machine with linear and puk kernel
function were used for classifying Parkinson data set with an accuracy of 90% [1]. In Artificial neural network method,70% of data wasused for training,and 30% for testing .Using this approach, 93.2% accuracy was achieved. The dataset consist of twenty three features [2]. By using maximumrelevance-minimum-redundancy criteria, features are selected on the basis of mutual information measures between the features [3]. In some cases,twenty three attributes are reduced to sixteen and 83.3% accuracy is achieved [4]. Various features subsets can be prepared and the subset which gives maximum accuracy is selected [5]. Genetic algorithm is used for feature selection. In genetic algorithm solutions are represented by chromosomes until acceptable results are obtained. Crossover and mutation process is done to get new chromosomes. With genetic algorithm for feature selection and support vector machine for classification, 94.5% accuracy is achieved [6]. When genetic algorithm with KNN classification method is applied, 98.2% performance is obtained[8].