|Parkinson's disease is a degenerative of central nervous system which causes the distruption of the nerve cells in the brain that will affect the movement of the sufferer. People affected by Parkinson's disease have low concentrations due to a lack of dopamine in the brain. Dopamine is a chemical in the body that serves as an introduction to a signal on the nerve. Early treatment using accurate method is needed in order to reduce the risk. A diagnosis method with data mining computation and machine learning is a good solution, in which data mining algorithm is used in classification of Parkinson's disease dataset. This study was aimed to compare of the performance result the Support Vector Machine (SVM) and linear regression algorithm in diagnosing Parkinson's disease. In order to increase the classification result, data mining algorithm can be combined with feature selection method. The comparison of both showed that, linear regression algorithm had higher classification than SVM by accuracy point of 88,7% and 87,7%. Meanwhile, after feature selection with Particle Swarm Optimization is done, the result of algorithm obtained higher score at the point of linear regression algorithm at 90,6% and SVM algorithm at 88,7%.
Keyword : Parkinson's disease; diagnosis; data mining; classification