Analytical study of Parkinson's diagnosis through classification techniques

Nerve Diseases are one of the most important health issues faced by a majority of the population of the world. They can range from as much as a small tooth sensitivity to more complex nervous diseases like Parkinson's disease or Parkinsonism. It is essential to have a frame work that can effectually recognize the prevalence of Parkinsonism in thousands of samples instantaneously. In this paper the potential of nine classification techniques is evaluated for prediction of Parkinsonism. Namely decision tree, naive Bayesian neural network, SVM, ANN, KNN. The proposed algorithm of SVM (support vector machine) employs in Parkinsonism prediction. Using medical profiles such as age, sex, blood pressure, muscle electric activity, EMG Evaluation, it can predict likeliness of patients getting Parkinsonism. Based on this, medical society takes interest in detecting and preventing the nerve disease. From the analysis, it has been proved that classification based techniques contribute high effectiveness and obtain high accuracy when compared to others.