医用电子鼻气体传感器阵列优化方法

Gas sensors array optimization for medical electronic noses

电子鼻气体传感器阵列存在冗余和相关,将对分类识别效果产生影响。提出并实现了一种新的医用电子鼻气体传感器阵列优化方法,利用改进遗传算法设定气体传感器的重要性系数,以去除传感器阵列的冗余和相关,提高系统的判别能力。新的方法分别用于五种乳癌特征气体和七种伤口病原菌检测,使正确识别率得到较大提高。

Since there are some redundancy and correlation in gas sensor array, which will affect the discriminating ability of e-nose （electronic-nose）, a new method for gas sensors array optimization for medical e-nose is designed and implemented. To reduce the redundancy and correlation of gas sensor array, and to enhance the discriminating ability of the e-nose system, the weight coefficient of each gas sensor is set by an improved genetic algorithm. The proposed new method is applied to five types of breast cancer biomarks and seven types of wound pathogens infections. The recognition accuracy can be significantly improved.