摘要
介绍了一种可以应用于气体识别领域的新的算法 -支持向量基算法 (SVM) ,并通过同常规的神经网络算法-BP算法进行实验对比 ,得到了 :SVM算法在数据样本不含噪声时可以得到和BP算法同样好的识别效果 ;在数据言本含有噪声时 ,该算法的识别效果相对BP算法具有明显的优势。从而证明了SVM算法在气体识别领域具有良好的研究价值和应用前景。
A new algorithm will be introduced which has not been paid enough attention to Support Vector Machines (SVM). Contrasting with BP algorithm which is very normal in the field of Neural Network, some useful conclusion can be gained during experiments as following: By using SVM, we can get the same impact as BP algorithm when the data do not have noises. But when the data has noises, using SVM will get better effect than using BP. The conclusion shows that SVM algorithm has well research value and applied foreground in the area of gas classification.
出处
《传感技术学报》
CAS
CSCD
北大核心
2005年第1期201-204,共4页
Chinese Journal of Sensors and Actuators
