Hybrid Classifiers for Gender Driven Emotion Recognition

Pravina Ladde

Abstract: Recognizing human emotions by registering speech signals is always an interesting field in Artificial Intelligence. This paper describes a system which makes 1] gender recognition first, to get apriori knowledge about the speaker.2] Then it uses hybrid of Hidden Markov Model [HMM] and Support Vector Machine [SVM] to make classification of emotions. This proposed system combines advantages of the classifiers to give more accurate results. Here HMM is used to model speech feature sequence i. e. this system is trained using HMM algorithm for considered emotion whereas SVM is used to make decision i. e. for classification. Literature survey and past results indicates that Gender Recognition apriori knowledge and use of hybrid HMM-SVM algorithms will considerably increase accuracy of emotion recognition.

Keywords: Hybrid Classifiers, Multiple classifier systems, Hidden Markov model, Support vector machine

How to Cite?: Pravina Ladde, "Hybrid Classifiers for Gender Driven Emotion Recognition", Volume 4 Issue 1, January 2015, International Journal of Science and Research (IJSR), Pages: 1236-1239, https://www.ijsr.net/getabstract.php?paperid=SUB15400, DOI: https://dx.doi.org/10.21275/SUB15400