Study of an Improved Online Least Squares Support Vector Machine Algorithm and Its Application in Gas Prediction

The paper studied on gas prediction problem. According to traditional prediction methods on coal mine safety being offline without dynamic prediction function and the shortcomings in the traditional online learning with least squares support vector machine (LS-SVM), this paper gave an improved online prediction algorithm of LS-SVM. This algorithm was used in gas prediction of some coal mine. By comparing with the actual data and other relative algorithms, the paper proved effect of the algorithm.