New Speech Emotion Recognition Analysis to Increase Accuracy Using SVM over Decision Tree

Abstract

Aim: To enhance the accuracy in classifying human speech based on emotions using Novel Support Vector Machine and Decision Tree. Materials and Methods: This study contains 2 groups i.e  Novel Support Vector Machine(SVM) and Decision Tree(DT). Each group consists of a sample size of 30 and the study parameters include alpha value 0.05, beta value 0.2, and the power value 0.8. Their accuracies are compared with each other using different sample sizes also. Results: The Novel Support Vector Machine is 80.67% more accurate than the Decision Tree of 70.37% in classifying the speech emotion of humans value of two tailed tests is 0.001 (p<0.5). Discussion and Conclusion:The Novel SVM model is significantly better than the DT in identifying human emotion via speech. It can be also considered as a better option for the classification of speech emotion.