Feature Selection Based Twin-Support Vector Machine for the Diagnosis of Parkinson’s Disease

With growing number of ageing population, Parkinson’s disease has become a serious problem to huge fraction of people above 60. The disease severely affects the motor system and can lead to death of the patients. There is no cure available for the disease. The symptoms on motor system is seen very late which leads into difficulty in management of the disease. There is no cure for the disease which makes it more difficult when the disease is diagnosed later. For early stages of Parkinson’s disease, there are some medications to improve the symptoms. There are certain symptoms like slurred speech, problems in utterances, etc. which are seen earlier. These symptoms can be leveraged to diagnose the disease in its earlier stages. Recently, computers and machine learning algorithms have been widely used in diagnosis of various diseases. Speech attributes can be analyzed using machine learning algorithms to build predictive models for detection of Parkinson’s disease. In this paper, twin-support vector machine (TSVM) based on feature selection technique has been discussed along with other ML techniques for the early diagnosis of Parkinson’s disease. Feature selection based TSVM has shown promising results in classification of Parkinson’s disease patients from healthy subjects.