基于KPCA-LSSVM的矿井工作面动力环境安全评价模型

为实现对采前工作面所处动力环境的客观、准确评价，选取9个直接影响工作面动力环境的指标因素构建安全评价指标体系，建立基于核主成分分析(KPCA)和最小二乘支持向量机(LSSVM)的工作面动力环境多因素耦合安全评价智能模型。首先根据KPCA理论对评价指标施行简约化处理，剔除冗余信息，得出6个简约后的评价指标并输入LSSVM模型中训练学习，最后得到评价模型。选取从平顶山矿区和大同矿区搜集到的30组工作面历史数据，按照20∶10的比例对模型进行训练和测试，并将测试结果与其他四种模型结果进行了对比，结果表明:KPCA方法可有效减少数据信息冗余，利用KPCA优化的LSSVM模型可准确评价工作面动力环境，误判率为0。

Safety evaluation model of dynamic environment in working face of mine base on KPCA-LSSVM

In order to evaluate the dynamic environment in working face before mining objectively and accurately, a safety evaluation index system was constructed by selecting nine indexes that influenced the dynamic environment in working face directly. A multi-factor coupling intelligent model of safety evaluation on dynamic environment in working face was built based on KPCA-LSSVM. Firstly, the reduction processing of evaluation indexes was carried out according to KPCA theory, and the redundant information was eliminated. Secondly, six evaluation in-dexes were obtained after reduction and input into LSSVM model to train and learn. Finally, the evaluation model was established. By selecting thirty groups of historical data in working faces collected from Pingdingshan mining area and Datong mining area, the model was trained and tested according to the proportion of 20:10, and the test results were compared with the results of other 4 models. The results showed that KPCA method can reduce the data information redundancy effectively, and the LSSVM model after optimization of KPCA can evaluate the dy-namic environment in working face accurately with an error rate of zero.