基于AE-CLSSA-ELM的煤与瓦斯突出危险性预测模型

Prediction model of coal and gas outburst risk based on AE-CLSSA-ELM

为了有效挖掘煤与瓦斯突出数据的非线性关系,提高煤与瓦斯突出危险性预测精度,提出1种基于自动编码器(AE)-改进麻雀搜索算法(CLSSA)-极限学习机(ELM)的预测模型。首先,在分析煤与瓦斯突出影响指标之间相关性的基础上,采用AE算法提取特征,降低数据复杂度;然后,基于麻雀搜索算法(SSA),引入Tent混沌映射和Levy飞行策略改进设计CLSSA;最后,利用CLSSA优选ELM的输入层权值和隐藏层阈值,构建煤与瓦斯突出预测模型对AE降维后的数据训练、测试,并与其他模型对比。研究结果表明:经AE特征提取后,ELM预测准确率提高了11%,且各类的错判数得到减少;基于AE-CLSSA-ELM的煤与瓦斯突出预测模型准确率为98.5%,F1值为97.87%,预测效果优于其他对比模型。研究结果可为煤与瓦斯突出事故的防范提供参考。

In order to effectively mine the nonlinear relationship in coal and gas outburst data and improve the prediction accuracy of coal and gas outburst risk,a prediction model based on the auto-encoder(AE),improved sparrow search algorithm(CLSSA)and extreme learning machine(ELM)was proposed.Firstly,based on the analysis of the correlation between the influencing factors of coal and gas outburst,the AE algorithm was used to extract features to reduce the data complexity.Then,based on the sparrow search algorithm(SSA),the Tent chaotic map and Levy flight strategy were introduced to improve and design CLSSA.Finally,the CLSSA was used to optimize the input layer weights and hidden layer thresholds of ELM,and a prediction model of coal and gas outburst was constructed to train and test the data after AE reduction dimension,and the prediction effect was compared with other models.The results showed that after AE feature extraction,the prediction accuracy of ELM was improved by 11%,and the number of wrong judgment of each subcategory was reduced.The accuracy and F 1 value of the prediction model of coal and gas outburst based on AE-CLSSA-ELM were 98.5%and 97.87%respectively,which were better than other comparison models.The results can provide reference for the prevention of coal and gas outburst accidents.