论文部分内容阅读
为消除人体穴位的阻抗信号特征集中存在的冗余和不相关分量的问题,提出了一种基于遗传算法的人体穴位阻抗特征子集选择与优化算法.通过分析穴位阻抗信号的自回归(AR)模型谱图建立了穴位原始特征样本集,利用类内-类间距离判据构造遗传算法的适应度函数并改进遗传算法的特征优化算子.经人体穴位的电阻抗特征选择与优化实例分析表明:该方法具有较好的寻优性能和适应度稳定,在不增加原始信息的情况下,能够有效地减少分类识别的特征数和提高信号识别的准确率,且将穴位阻抗特征的平均状态辨识率提高9%左右.
In order to eliminate the redundant and uncorrelated components of the impedance signal characteristic of human acupoints, a genetic algorithm based algorithm for selecting and optimizing the feature points of acupoint impedance is proposed. By analyzing the autoregressive (AR) The original feature sample set of acupoints was established based on the model spectrum, and the fitness function of genetic algorithm was constructed by using the intra-class-distance distance criterion and the feature optimization operator of genetic algorithm was improved.According to the selection of electro-impedance characteristics of human acupoints and the optimization example analysis : This method has good performance of optimization and stability of fitness. Without adding original information, the method can effectively reduce the number of features of classification and recognition and improve the accuracy of signal recognition. And the average state identification of impedance characteristics of points Rate increased by about 9%.