Diagnosis Of Schizophrenia Based On Support Vector Machine And Cognitive Function Behavior Characterization

Schizophrenia is a mental disorder,including mental and perceptual distortions,emotional apathy and behavioral disorders.For a long time,schizophrenia is very destructive to the life and work of the patients,and the migration and repetition of the course of the disease have brought serious neurocognitive impairment to the patients.This indicates that cognitive function should be used as an important dimension of the disease to introduce clinical practice in order to provide great clinical application value to the early identification and diagnosis of schizophrenia.In addition to clinical scale evaluation and imaging analysis,there is still a lack of objective and convenient methods for the diagnosis of schizophrenia.Therefore,the research on the auxiliary diagnosis technology of schizophrenia has become a hot spot.With the rapid development of AI wave,the application of machine learning to realize the auxiliary diagnosis of schizophrenia will have high clinical application value and market prospect.In order to solve the existing problems,this study uses an experimental system to evaluate cognitive disorders in schizophrenic patients,combining cognitive psychology experiments with machine learning methods to objectively identify schizophrenic patients.A total of 100 subjects were included in this study,including 55 normal subjects and 45 patients.By analyzing the reaction time and error rate,the behavioral response characteristics of schizophrenic patients were studied.Firstly,feature selection algorithm is applied to extract the key features of the original features of 13 experimental paradigms.Secondly,the particle swarm optimization classification algorithm is used to classify the key feature combinations of different experiments.Finally,an experimental paradigm set that can accurately classify schizophrenia is selected.The results show that the classification accuracy of classification model E established by all key feature combinations extracted from the five experiments reaches 99%,and the AUC value is 0.97.The classification results show that the key feature combination of these experimental tasks can effectively distinguish schizophrenia and can be used as an objective auxiliary diagnostic method to improve the accuracy of the doctor’s diagnosis.The main research work of this paper includes:(1)Cognitive impairment is common in people with schizophrenia.Therefore,thirteen experimental paradigms of behavioral cognitive psychology were selected to achieve the purpose of the study.(2)At first,a preliminary statistical analysis and preprocessing of the data from the 13 experiments is conducted,and find that there are obvious differences in the reaction time between the patient group and the healthy group,and then construct the corresponding characteristic variables according to the original data.Relief algorithm is used to extract the key information from the behavioral data and to mine the effective information from the data.(3)Through the comparison of the performance of various parameters optimization learning algorithm,the appropriate classifier is selected.The classification of feature subsets is studied for a variety of feature combinations to obtain the final classification results.By comparing and analyzing the experimental results,the experimental paradigm with higher classification accuracy is selected from the result set,and various experimental combinations are considered synthetically in order to achieve the optimal classification accuracy.(4)The above experiment process needs to load data and save data and call different algorithms frequently in MATLAB.Taking into account the convenience of personnel operations and saving time,this study proposes the use of MATLAB programming language to integrate the above data analysis process,and designs a schizophrenia disease diagnosis system,making data analysis and category diagnosis efficient and rapid.