Research And Implementation Of Parkinson Disease Recognition System Based On Speech Recognition

With the coming of the aging society and the steady improvement of the living standards of people,the health of the people is not in good condition,therefore more and more people pay more attention to their health.Parkinson’s disease is one of the fatal diseases having a great influence on the happy life of the elderly,so it is necessary to prevent it.But the traditional method of detecting Parkinson’s disease need professional to implement,which has long detection period,complex operation and cannot satisfy the demand of the adapting to the fast detection.Because the implement of detecting Parkinson’s disease based on voice data has been proved to be an effective method recently,but at present,the relative researches do not take the preprocess of sample and the Integrated learning into consideration well,which causes the poor accuracy of classification,the unsteadiness and the sample has an influence on classifier.At present,the system of detecting Parkinson’s disease has problems of complex operation and poor algorithm,which cannot satisfy the demand of experimenters and hinders the progress of experiments.So the article puts forward a prediction model for detecting Parkinson’s disease using phonetic amplitude,pitch frequency,Short-time zero crossing rate and age of disease as four features,aiming at Inaccurate prediction results with less feature for detecting Parkinson’s disease,and improve the accuracy rate of the prediction model by adjusting the weight of the characteristic parameters of experiment constantly,and develops a system of detecting Parkinson’s disease based on Mobile terminal Using android language design and MATLAB development.The system can run-up the experiment with the features of steady operation,convenience,high accuracy,and friendly interface.In the paper,main works are as follow:Firstly,the paper introduces Parkinson’s disease and the background and meaning of subject briefly,introduce its development status at home and abroad meanwhile;then,introduce many relative technologies that need to be used in developing the system according to analysis the demand of software system;then introduce the key part of the paper,that is the detail of the core algorithm of the system,which is divided into the following parts:(1)The design of interface.I make the interface design scheme of the system according to the design principles and goals of the system.(2)Module design and Implementation.I determinate the solution by analyzing the several core problems in implementing the system.And state the design ideas and code of each module,and divide the functions of the system into different module according to analyzing the functions and experimental procedures of the system.(3)The research of classification algorithm.The paper puts forward a classification algorithm that combines random forest and sample repeating editing.The prediction model has a good classification performance based on the two classification algorithm,so the paper uses the prediction model to process on voice data.The algorithm the paper puts forward has a better result comparing with other different algorithms.Finally,it shows that the software system meets the design requirements and can be used into the practical application according to a serious testing for the system of detecting Parkinson’s disease.