冲击地压Fisher判别分析预测

Prediction of rock burst based on Fisher discriminant analysis

为对冲击地压危险性进行准确地预测预报.在综合分析冲击地压多种影响因素的基础上,采用Fisher判别分析方法,结合文献提供的张集矿历年冲击地压24组相关数据实例,建立了冲击地压预测预报的Fisher判别分析模型,模型回代预测准确率为100%,误判率为0.将模型应用到张集煤矿冲击地压分级预测问题中,并与支持向量机方法、BP网络进行比较,研究结果表明:Fisher判别分析模型具有较高的可靠性,判别效率及准确率高,能对冲击地压进行有效预测.

In order to predict the rock burst hazard areas accurately, this paper applied Fisher discriminant analysis based on a comprehensive analysis of many factors of rock burst. Provided with 24 groups of related data instances of rock burst from each year offered by Zhangji mine, this study established the Fisher discriminant analysis model of rock burst prediction, whose back substitution accurate rate was 100% while false rate is 0%. This paper applied this model to the problem of rock burst prediction in classification of Zhangji mine, and then compared it with the approach of support vector machine and BP network. The results show that the Fisher discriminant analysis model has a relatively high reliability, high determining efficiency and accuracy, which can be used to predict rock burst effectively.