基于未确知测度理论,建立回采工作面瓦斯涌出量的均值属性测度聚类预测模型。模型以样本均值为聚类中心,以熵权确定评价指标测度的权重,通过计算样本熵权综合测度与所属类别目标均值乘积之和获得瓦斯涌出量的预测值。利用实测的20组数据作为训练样本建立预测模型,并对校本进行逐一检验。研究结果表明:瓦斯涌出量预测值与实测值的相对误差不超过5%,能满足工程需要;与支持向量机(SVM)工具的验算相比,易于为现场的工程技术人员所掌握。
Based on the theory of unascertained measure, the mean attributed measurement clustering model for coal face gas emission prediction was established. In this model, the sample mean was set as the cluster center, and the weight index of evaluation parameters were determined by entropy. Through calculating the sum of the product of sample entropy weight comprehensive measurement and target mean of classification that the sample belonged to, the gas emission prediction value was obtained. 20 groups of data gotten from a mine as training samples were used to build forecasting model, and the model was tested by the data. The results show that the relative error between predicted emissions and the measured values is less than 5%. The model can meet the need of engineering, and compared with the support vector machine (SVM) calculation tool, it is easy to be masterd by field engineering and technical personnel.