采用计算机进行情绪判断对实现人工智能、人机交互及智能计算等具有重要意义。本文在深入学习和研究脑电信号分析和处理的各种算法基础上，进行了对基于ELM的情绪分类模型研究。本文采用近似熵和小波能量熵算法生成三种不同的脑电信号特征，对ELM分类器进行训练，同时与BP算法、GRNN和PNN算法进行比较。实验表明，ELM极限学习机算法分类效果最好，其识别率达到87.25%。%The use of computer for emotional judgment is of great significance to the realization of artificial intelligence, human-computer interaction and intelligent computing. Based on the in-depth study and research of the processing algorithms of the EEG signal analysis, this paper studies the sentiment classification model based on ELM. In this paper, approximate entropy and wavelet energy entropy algorithm are used to generate three different features of EEG signals to train the ELM classifier. At the same time, it is compared with BP algorithm, GRNN and PNN algorithm. Experiments show that ELM limit learning machine algorithm classification effect is the best, the recognition rate is 87.25%.
展开▼
