Research on voiceprint recognition based on weighted clustering recognition SVM algorithm
Yang Wu, Lihong Xu, Yandong Chen, Xueyang Zhang · 2017
Support vector machine (SVM) algorithm received much attention in the research of voiceprint recognition, especially for small sample datasets. However, with the increase of recognition number and speech features number, the rate of model training and recognition is significantly reduced. In order to solve the problem, a new weighted clustering algorithm is proposed, which use “one to one” SVM model to reduce the model complexity. In training model stage, this paper adopts sequential minimal optimization (SMO) algorithm and quadratic programming and heuristic method to select variables speeding up the model training. In identification stage, use k-means and linear programming combined way to get each person's recognition score. Based on 100 TIMIT speech data, experiments show that the recognition rate is greatly improved compared with the traditional method when the number of recognition and training eigenvector reaches certain conditions.