An Expert System for Identification of Key Factors of Parkinson's Disease: B-TDS-PD

Abstract

Due to modern means of living and falling to take care of the health an individual may suffer a severe disease known as Parkinson's Disease. It is a disorder which affects a small region of the brain which controls movement, physical posture and also the emotion of an individual. PD generally occurs in elderly people but with time it has become a chronic disease affecting people of any age group. Keeping in view of its severity, many researches have been done to control the disease and if an expert system which can identify the major risk factors of PD than it would be of great achievement. Hence this paper proposes an expert system named Balanced- Transparent Decision System for-PD (B-TDS-PD) which identifies the major risk factors of PD by the rules generated by Decision Tree. B-TDS-PD encompasses of 5 stages: Preprocessing, Rule Creation, Rule Picking, Rule Pruning and Integrating and Key Factor Identification. The PD data can be imbalance in nature means the ratio of positive class to negative class may be very high, so to balance preprocessing technique is used. In the Rule Creation step, the decision rules are generated from Decision Tree and in the Rule picking step the transparent decision rules are selected. Later in the Rule Pruning and Integrating step, the redundant and the incompetent rules from the rule set are removed and integrates to a single rule. Finally, in the last step the key factor(s) of PD are identified. For experimentation the Parkinson's Disease Speech dataset from UCI is considered.