Classification of Roadway Surrounding Rock Stability Based on Random Forest

In the article, rock strength, depth, fracture development, width of roadway, thickness ratio of immediate roof and coal seam, thickness of the loosen zone are selected as sample variables based on roadway stability influencing factors. Through related-data of 35 roadways collected in some coal mines, roadway stability classification model is established by random forest, and the prediction result is compared with the decision tree, BP neural networks and support vector machines model. The results show that random forest model can relatively and effectively determine the stability of roadway with low false rate and high prediction accuracy.