Research on Medical Voice Recognition and Restoration Based on Support Vector Machine Algorithm

The essence of medical voice recognition and restoration is to digitally process the collected human voice signals to obtain various voice-related features, so as to evaluate the voice quality. With the rapid development of pattern recognition, the classification of normal and pathological voices can no longer meet the current clinical needs, but gradually turns to the classification between different degrees of pathology. In this paper, the parameter fusion is combined with the classifier SVM (Support vector machine), and the weights are matched by the scores of the output of SVM. Finally, a variety of classifiers are used to identify and repair medical voice. The research results show that the accuracy is 0.974 and AUC is 0.95 by using the feature parameter fusion algorithm of SVM weight matching in this paper, which proves the effectiveness of this algorithm.