Characteristic analysis and preprocessing of mine gas monitoring data

In view of characteristics of abnormal data, missing data and noisy data of mine gas monitoring data, a preprocessing method of gas monitoring data was proposed. Abnormal data is replaced by use of moving average line processing method or auto regressive model processing method, missing data is filled by employing cubic exponential smoothing method and data denoising is processed though wavelet soft threshold method. The case analysis shows that the method can eliminate interference of abnormal data, ensure integrity of monitoring data and smooth characteristic curve of monitoring data without changing statistical characteristics of gas monitoring data.