Development Of Speech Feature Extraction Kit Of Parkinson’s Disease And Research Of Three-way Decisions Diagnosis

Parkinson’s disease(PD) is a common neurodegenerative disease of the central nervous system, its condition is causes loss of some or all of the motion response,includeing speech, behavior, mental process and other important functions.. But, if can be diagnosed in the early onset of disease and treatment in time, it can efficiently alleviate the deterioration of disease, The language barrier is one of the early onsets of Parkinson’s disease, The resaerch in Parkinson’s disease diagnosis based on the language barrier is very popular in recent years.Firstly,in the light of the speech disorder of Parkinson’s characteristics, this paper collects some of the effective algorithm in current,and picks up some features those can reflect the differences between the patients of Parkinson’Disease’disease and healthy pepole vie compare,then develops a Feature-extraction software whitch is run on windows platform via combine these algorithm with practical.This software is easy to use,so make it easier for unprofessional people with no professional knowladge in signal processing to use it.Then gather speech from patients of Parkinson’Disease’disease and healthy people in practice,and extract featurs by this software kit save the extracted data as train database.Secondly,study The theory of three-way decision,understand the defference between the traditional two-way decisions and three-way decision and post a kind of structure of classifier based on this theory—three-way decision classifier,and build a model on matlab.Then try to optimize it from several aspects,and test it via savel UCI data sets.Compare the variation in property in those differsnt optimization method,and find out the best optimization method.Finally,extract features form the actual speech by feature extraction software,and classify both data sets,including the self-extracted data set and Parkinson’Disease’disease data set in UCI by optimized three-way decision classifier,to achieve the complete function of the voice diagnosis of Parkinson’s disease.