基于极大似然估计的logistic回归模型在煤与瓦斯突出危险等级评价中的应用

Application of the Logistic Regression Model in Coal Gas Outburst Dangerous Grade Evaluation Based on Maximum Likelihood Estimation

为了提高煤与瓦斯突出危险等级预测的准确性和增强预测预报方法的实用性,选取开采深度、瓦斯压力等五个关键影响因素作为煤与瓦斯突出的评判指标,建立了煤与瓦斯突出多元logistic回归非线性概率评级模型,对模型参数的确定采用了极大似然函数估计法,并利用MATLAB遗传算法工具箱的主要函数ga()求出模型系数的初始值,再用搜索工具箱中的fminsearch函数求出模型系数的最终值。通过对16组样本的学习和对6组样本的预测,验证了该模型在煤与瓦斯突出危险等级判别中的有效性、准确性。另外,该模型结构简单,计算量小、实用性强,易于程序实现,有借鉴价值。

In order to improve the accuracy of prediction of coal and gas outburst dangerous grade and enhance the usefulness of prediction methods, selecting five key factors of mining depth, gas pressure, etc. , as evaluation index of coal and gas outburst, set up multi- ple logistic regression nonlinear probability rating , the determination of the model parameters by using thethe method of maximum likeli- hood estimation, and use the main function ga（ ） in MATLAB genetic Algorithm Toolbox to obtain the initial value of the model coeffi- cients, and then use search function fminsearch（ ） in search toolbox for a final value of model coefficients. Through the 16 groups of sample learning and prediction of 6 groups of samples, verify the validity, accuracy of the model in the coal and gas outburst dangerous grade discrimination . In addition, the model has the advantages of simple structure, small amount of calculation, strong practicability, easy to program implementation, there is reference value.