ZHAO Meicheng, HE Anmin, QU Shijia. Research on time series prediction method of gas data on fully mechanized mining face[J]. Journal of Mine Automation, 2019, 45(6): 80-85. DOI: 10.13272/j.issn.1671-251x.17421
Citation: ZHAO Meicheng, HE Anmin, QU Shijia. Research on time series prediction method of gas data on fully mechanized mining face[J]. Journal of Mine Automation, 2019, 45(6): 80-85. DOI: 10.13272/j.issn.1671-251x.17421

Research on time series prediction method of gas data on fully mechanized mining face

  • In view of problems of complex algorithm and short prediction step length of existing gas concentration prediction methods based on time series, a time series prediction method of gas data on fully mechanized mining face based on ARIMA+GARCH combination model was proposed according to randomness and timing of historical monitoring data of gas concentration. Firstly, an ARIMA prediction model is established, and then the gas concentration data is smoothed and the parameter estimation of the model is determined. After the reliability of the prediction model is passed the test, the GARCH model is used to simulate fitting residual error of ARIMA for the mean regression problem existed in the prediction process of ARIMA model. The simulated results are used as the noise term of prediction in ARIMA to optimize the prediction result. The test results show that the gas concentration prediction methods based on ARIMA+GARCH combined model can reflect the change trend of the true value of gas concentration, and the mean absolute deviation, mean absolute percent error, standard deviation error and mean squared error are all small, and has high prediction accuracy.
  • loading

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return