Design and Analysis of Data Mining Based Prediction Model forParkinson’s disease

Abstract: Purpose: The purpose of this research paper is to develop a prediction model for Parkinson’s disease. There are many symptoms that lead to Parkinson’s disease such ageenvironmentalfactor, trembling in the legs, arms, hands, impaired speech articulation and production difficulties. In this research paper speech articulation of Parkinson’s disease affected people is considered for model formation and analyzes the model based on the symptom of disease. Methods: In proposed prediction model tree based classification model decision tree, ID3 and decision stumps are used for training and testing the effectiveness of proposed prediction model. Here we also applied K –fold cross validation technique for true prediction so that each record is sued for training and testing. Results : In proposed model decision tree based our prediction model provide accuracy 85.08%, classification error 14.92%, ID3 provide accuracy 75.33% ,classification error 24.67% and decision stumps based model proved accuracy 83.55% and classification error 16.45%. Conclusion:Proposed model based on Decision tree provide best result in comparison to other in terms of parameters accuracy and classification error. Keyword:Parkinsons, Data mining, Decision tree, ID3, Decision Stumps.