Rock burst intensity grading prediction based on the combination of PCA，CBLOF and SVMSMOTE algorithms
YAO Nan1，2，Zhang Yili1，LIU Yang1，2，YE Yicheng1，2，LUO Xiaobing3，FENG Jie3，HUANG Zhaoyun4
(1. School of Resource and Environmental Engineering，Wuhan University of Science and Technology，Wuhan，Hubei 430081，China；2. Hubei Key Laboratory for Efficient Utilization and Agglomeration of Metallergic Mineral Resource，Wuhan，
Hubei 430081，China；3. Hubei Provincial Emergency Rescue Center，Wuhan，Hubei 430000，China；
4. Hubei Jingshen Safety Technology Co.，Ltd.，Yichang，Hubei 443000，China)
Abstract：To reduce the correlation between feature parameters in rock burst sample data，remove outliers，and balance the number of various rock burst levels in the dataset，a method combining PCA(principal compo-nent analysis)，CBLOF(cluster-based local outlier factor)，and SVMSMOTE(support vector machine synthetic minority over-sampling technique) is proposed to enhance the quality of the rock burst database. Initially，343 rock burst cases from both domestic and international sources were collected to build the original rock burst da-taset. PCA is employed for dimensionality reduction，CBLOF is used to eliminate outliers within each rock burst level，and SVMSMOTE synthesized new minority samples near the boundaries of each rock burst level. The processed and original rock burst databases were then trained using six different machine learning models to vali-date the effectiveness of the PCA，CBLOF，and SVMSMOTE combination. Results show that the accuracy of models trained on the processed database improved significantly：AdaBoost by 29%，CatBoost by 28.5%，LightGBM by 34%，GradientBoosting by 28%，ExtraTrees by 26.5%，and RandomForest by 24%，compared to models trained on the original database. Therefore，processing the rock burst database using the combined PCA，CBLOF，and SVMSMOTE algorithms effectively enhances the quality of the database and improves the performance of machine learning prediction models.
姚囝1，2，张义礼1，刘 洋1，2，叶义成1，2，骆效兵3，冯 杰3，黄兆云4. 基于PCA，CBLOF和SVMSMOTE算法组合的岩爆烈度等级分级预测[J]. 岩石力学与工程学报, 0, (): 967-967.
YAO Nan1，2，Zhang Yili1，LIU Yang1，2，YE Yicheng1，2，LUO Xiaobing3，FENG Jie3，HUANG Zhaoyun4. Rock burst intensity grading prediction based on the combination of PCA，CBLOF and SVMSMOTE algorithms. , 0, (): 967-967.
