自适应特征选择的煤矿突水预测模型研究

Research on Coal Mine Water Inrush Forecasting Model Based on Adaptive Feature Selection

摘要: 针对煤矿突水影响因素复杂难以预测的问题，通过理论分析，构建影响煤矿突水因素的指标体系，并针对收集的相关数据，提出基于遗传算法的自适应预测模型的特征选择算法。实验结果表明：相比于稳定性特征选择算法和递归特征消除算法，自适应特征选择算法可以更好地提高模型预测准确率，适应性更强。

Abstract: In view of the complex and unpredictable influencing factors of coal mine water inrush, the index system of influencing factors of coal mine water inrush is constructed through theoretical analysis. And according to the relevant data collected, the feature selection algorithm of adaptive prediction model based on genetic algorithm is proposed. The results show that compared with stability feature selection algorithm and recursive feature elimination algorithm, adaptive feature selection algorithm can improve the accuracy of model prediction better and has stronger adaptability.