Early Diagnosis of Parkinson's Disease with Speech Pronunciation Features Based on XGBoost Model

Speech disorder is one of the early symptoms of Parkinson's disease. In order to accurately diagnose the Parkinson's disease in early time, we propose a Parkinson's disease diagnosis method based on machine learning. XGBoost algorithm was applied to detect and classify the speech signals of Parkinson's disease patients. The performance of this method has been evaluated on several metrics based on the UCI Parkinson database, and in experiments, support vector machines, random forests, and neural networks have also been used to detect and classify the speech signal collected from Parkinson's patients. Experimental results show that the performance of XGBoost algorithm is better than other algorithms, the accuracy of Parkinson's disease diagnosis is 96%, precision is 100%, AUC is 0.97, F1_ Score is 0.97.It thus contributes to the better understanding of the Parkinson disease and possibly the further analyze of the speech feature.