Support vector machines approaches and its application to speaker identification
S. Zribi Boujelbene, D. Ben Ayed Mezghani, Noureddine Ellouze · 2009
This paper proposes a classification approach that incorporates the statistical methods GMM and support vector machines. The proposed GMM-SVM system is presented and experimentally evaluated on text independent speaker identification. Our results prove that the combination approach GMM-SVM is significantly superior than SVM approach. We report improvements of 85,37% amelioration in identification rate compared to the SVM identification rate.