The Diagnosis of Parkinson's Disease Based on Gait, Speech Analysis and Machine Learning Techniques

Abstract

Parkinson's disease (PD) is a long-term degenerative disorder of the central nervous system. The common symptoms are tremor, rigidity, slowness of movement, and difficulty with walking at early stages. Currently, PD can't be cured. And there are not really effective methods to diagnose it. However, machine learning is a new way for the diagnosis of PD. It can build a model from PD patients' dataset, which can help classify PD and healthy people. In this review, the applications of machine learning for PD diagnosis by algorithms and data are analyzed. Several machine learning classifiers are briefly introduced, including artificial neural network (ANN), support vector machine (SVM), Naive Bayes (NB), K-Nearest Neighbor (k-NN). Next, the basis of gait analysis is introduced, including gait circle and gait data, and then, each step of the machine learning processing is focused on. Two ways are concentrated to analyze speech signals - support vector machine (SVM) and artificial neural network (ANN). This review presents that machine learning has good performances for the diagnosis of PD. However, it can only be a diagnosis tool to help doctors because of its limited generalization. In the future, people should explore more effective algorithms with better generalization.