Detecting Parkinson's Disease Using Gait Analysis with Particle Swarm Optimization

Gait analysis is the study of human movements by analyzing temporal and spatial gait features. Research has shown that Parkinson's disease can degenerate human mobility, thereby causing afflicted individuals to behave differently in terms of gait characteristics. In this work, we propose an optimized method that assists us in better distinguishing people with Parkinson's disease from normal subjects. The spatial-temporal gait features are extracted by using a real U-shaped pressure-sensitive gait-sensing walkway. After pre-processing optimizations, including nondimensionalization and normalization of the raw features, we feed the features to an SVM classifier for training. The Particle Swarm Optimization algorithm is adopted to optimize the classification model. Experimental results show that the optimized method outperforms its predecessor by improving the accuracy from 87.12% to 95.66%, which shows the effectiveness of our proposed method in detecting Parkinson's Disease patients.