基于进化策略的生理信号情感识别

Recognition of emotion in physiological signals using evolutionary strategies

针对生理信号的情感识别问题,采用进化策略方法对生理信号进行特征选择,利用智能优化算法的计算复杂度低和收敛速度快等特点,并结合使用近邻法进行分类,有效地解决了生理信号特征组合优化问题,筛选出一定的特征子集来表示相应的人类情感状态.实验仿真表明,该方法可以得到有效的特征组合来进行生理信号的情感状态识别.

A novel method was developed to perform feature selection in order to recognize emotion in physiological signals. By combining the near neighbor classifier and the intelligent optimization algorithm, combinatorial optimization problems in physiological signals can be effectively solved and a series of important features can be selected that represent a human＇s emotional status. Simulations showed that it effectively uses features so extracted from experimental data for recognition of emotional states.