The prediction model of gas emission is established based on partial correlation analysis and support vector regression (SVR) in order to accurately predict gas emission of working face under the condition of small samples. Not only are the problems of small samples and nonlinear prediction effectively resolved by applying SVR, but also the main control factors of gas emission are selected by applying partial correlation analysis method, which can reduce variables space dimension of the model to improve prediction accuracy. Through empirical analysis, the superiority of the model is proved by prediction results that are quite close to the measured values.
Yang, Li and Liu, Chengcheng
"Prediction of Gas Emission Based on Partial Correlation Analysis and SVR,"
Applied Mathematics & Information Sciences: Vol. 07:
5, Article 3.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol07/iss5/3