Abstract: For the problem of emotion recognition, genetic algorithm based on simulated-annealing method, max-min ant colony algorithm and particle swarm algorithm were used for feature selection, and combined with Fisher linear classifier to recognize six emotions: joy, surprise, disgust, grief, anger and fear, it has obtained higher recognition rate. Effective feature subset which can identify the emotion recognition system model with better perfomfance was found, and the recognition system was established with forecasting ability of six emotions.

Key words: Emotion recognition, Feature selection, Intelligent algorithm