武汉工程职业技术学院学报2025，Vol.37Issue(1)：29-34,59,7.
基于GCN-LSTM架构的煤矿瓦斯预测方法
A Coal Mine Gas Prediction Method Based on GCN-LSTM Architecture
摘要
Abstract
To address the issue of low prediction accuracy in coal mine gas concentration data,a spatio-temporal neural network model combining Graph Convolutional Network(GCN)and Long Short-Term Memory(LSTM)networks is proposed.This model enhances prediction accuracy by employing GCN and LSTM layers to construct spatiotemporal blocks that effectively capture the spatial and temporal features of coal mine gas data,thereby improving the model's capacity to identify complex data patterns and gener-ate more accurate forecasts.Experimental results demonstrate that the proposed method outperforms GRU and LSTM in terms of fitting accuracy,with Root Mean Square Error(RMS E),Mean Absolute Error(MAE),and Coefficient of Determination(R2)values of 0.2266,0.1875,and 0.8206,respectively,sur-passing those achieved by GRU and LSTM models.These findings confirm the effectiveness of the pro-posed method in enhancing coal mine gas concentration predictions.关键词
图神经网络/长短期记忆网络/无监督学习/煤矿瓦斯预测Key words
graph neural networks/long short-term memory networks/unsupervised learning/coal mine gas prediction分类
矿山工程引用本文复制引用
高成,盛武..基于GCN-LSTM架构的煤矿瓦斯预测方法[J].武汉工程职业技术学院学报,2025,37(1):29-34,59,7.基金项目
安徽省自然科学基金项目(1808085MG212) （1808085MG212）
安徽理工大学2023年研究生创新基金项目(2023cx2162) （2023cx2162）
