Predicting Parkinson's Disease Using Different Features Based on Xgboost of Voice Data

Abstract

The purpose of this article is to determine Parkinson's disease (PD) is a neurotic condition characterized by the demise of nerve cells in the middle nervous system. The neurologic capacity is speech or voice recognition to predict if a body has been applied speech stored dataset of patients, to use an engine learning algorithm to analyze the sound patients to predict the PD patients that affect approximately 92 percent of patients, a voice problem issue, to work on dataset decision tree to predict the PD with maximum exactitude. It was the better pattern to utilize on the data with an accuracy of 90–96 percent (PD), a format speaks signals, to offer a better outcome from our patients who modify the old age system above the time of 66 years and it develops at a superior price till 2060. Several contemporary engine learning and pattern recognition techniques were used in this study to categorize or predict the risk of Parkinson's disease based on speech signal data. A number of classification approaches, including as K-NN, Decision Trees, and Neural Networks, are presented in this project, as well as some “Ensemble” Gradient boosting, which is an engine that learns reflux and grouping difficulty knowledge. This results in an ensemble of incapable divination patterns as a divination pattern. Within coming period, combining voice messages and some other medical information, our system will help clinicians in more accurately and swiftly identifying the PD subgroup from of the normal participants.