Detection and analysis of emotion recognition from speech signals using Decision Tree and comparing with Support Vector Machine

Abstract

The aim of the study is to detect and classify emotions from speech signals using Decision Tree (DT) classifiers compared to Support Vector Machine (SVM) classifiers. Emotions are detected from speech signals using a DT and SVM classifier. The unique aspect of this research is the selection of salient elements from speech signals. The total sample size is 40. In the classification of eight different emotions, the DT classifier has a 35.90 % accuracy, whereas the SVM has a 27.18 % accuracy. There is a significant difference between the two groups, since p < 0.05 in SPSS Statistical analysis. From this work, it is observed that the accuracy obtained in detecting human emotions using the DT classifier is significantly better than SVM.