Gas Concentration Prediction Based on SMOTER and Extreme Learning Machine
Gas disaster has always been an important factor restricting the safe productivity of coal mines.Accurately predict-ing the trend of gas concentration in a near future can effectively avoid coal safety accidents and provide reference information for preventing gas disaster and the safe productivity of coal mines.The gas concentration in most time is maintained at the normal level,which results in the imbalanced distribution of the original gas concentration data.This study proposes a gas concentration prediction method based on SMOTER and extreme learning machine(ELM).Firstly,the relevance function is used to separate the data set in-to rare value cases and normal value cases,and then SMOTER algorithm is used to generate rare value cases to process imbalance data set.Finally,the balanced data set is used to train the ELM gas concentration prediction model.The proposed method uses SMOTER algorithm to address the imbalance problem in data of regression task,which makes the gas concentration prediction mod-el can be accurately predicted on the rare value cases.Compared with the existing methods,the experimental results show that the prediction error of the proposed method is smaller.
gas concentration predictionsafe productivity of coal minesSMOTERextreme learning machineregression
国家科技期刊平台
NSTL
万方数据
