Emotional Analysis Using Multinomial Logistic Regression

Background: Emotions have been widely used in psychology and behavior sciences as they are important elements of human nature. We expect a machine to behave like a human, in this digital era. Since the machines do not understand the emotional state of the speaker easily, it is not very easy to get a natural interaction between machine and man. Yet, many researchers are working and progressing in speech recognition. In this paper we aim to identify emotions present in human beings through analysis of speech signals. Methods: We make use of machine learning algorithm to choose the best features which effectively influence the emotional states of speech. Findings: We aim to identify various emotions that a human being goes through during verbal communication. Applications: Computation on the chosen best features of speech signals help us identifying the speaker’s emotional state.