Speech emotion recognition system based on L1 regularized linear regression and decision fusion

摘要: This paper describes a speech emotion recognition system that is built for Audio Sub-Challenge of Audio/Visual Emotion Challenge (AVEC 2011). In this system, feature selection conducted via L1 regularized linear regression in which the norm weights minimized to find sparse weight vector. The features with approximately zero are removed create well-selected small set. A fusion scheme by combining strength from and Extreme learning machine (EML) based feedforward neural networks (NN) proposed classification. experiment results on SEMAINE database naturalistic dialogues distributed through AVEC 2011 presented.