基于四大集成学习的岩爆烈度分级预测

doi:10.13722/j.cnki.jrme.2022.0026

Abstract
Figure/Table
References (0)
Related Citation (15)

Download: PDF (748 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract In order to accurately predict rockburst disasters，the four integrated learning of Boosting，Bagging，Stacking，and Voting are applied to rockburst disaster prediction. The prediction performance of common machine learning algorithms，Boosting and Bagging is compared. The base model selection method of Stacking and Voting is proposed. First，275 groups of rockburst cases at home and abroad were collected to construct the original rockburst data set，statistical parameters of different rockburst grades in the original rockburst dataset were analyzed. TSNE(t-distributed Stochastic Neighbor Embedding) algorithm is used to visualization analysis the original rockburst data. The analysis shows that there are a large number of outliers in the original rockburst data set and the data is unbalanced. Secondly，Yeo-Johnson transformation and K-means SMOTE oversampling are used successively to normalize and balance the data to reduce the effects of outliers and data imbalance，respectively. The normalized and balanced rockburst data has enhanced separability. Then，Training prediction is made for 15 machine learning algorithms including general machine learning，Boosting and Bagging. The accuracy of macro average and Friedman statistical hypothesis test were used to compare the prediction performance of various models. Finally，a prediction method of rockburst intensity grading based on Stacking and Voting is proposed. The models with low accuracy or high similarity of prediction results were eliminated by using the exhaustive method. General machine learning，Boosting and Bagging algorithms retain the models with high precision and as different as possible as the base models for Stacking or Voting. The results show that the prediction performance of Boosting，Bagging，Stacking and Voting is mostly better than that of ordinary machine learning. The Stacking and Voting rockburst intensity classification prediction methods combining diversity and accuracy weights can effectively improve the rockburst prediction performance.

Key words： rock mechanics ensemble learning TSNE dimensionality reduction visualization data normalization and balance diversity and accuracy weights combined rockburst intensity classification prediction

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	TAN Wenkan1
	2
	HU Nanyan1
	2
	YE Yicheng1
	2
	WU Menglong1
	2
	HUANG Zhaoyun3
	WANG Xianhua4

Cite this article:

TAN Wenkan1,2,HU Nanyan1, et al. Rockburst intensity classification prediction based on four ensemble learning[J]. , 2022, 41(S2): 3250-3259.

URL:

http://rockmech.whrsm.ac.cn/EN/10.13722/j.cnki.jrme.2022.0026 OR http://rockmech.whrsm.ac.cn/EN/Y2022/V41/IS2/3250

[14]	赵洪波. 岩爆分类的支持向量机方法[J]. 岩土力学，2005，26(4)：642–644.(ZHAO Hongbo. Support vector machine method for rockburst classification[J]. Rock and Soil Mechanics，2005，26(4)：642–644.(in Chinese))
[10]	谢和平. 岩爆的分形特征及机制[J]. 岩石力学与工程学报，1993，12(1)：28–37.(XIE Heping. Fractal characteristics and mechanism of rockburst[J]. Chinese Journal of Rock Mechanics and Engineering，1993，12(1)：28–37.(in Chinese))
[32]	BREIMAN L. Bagging predictors[J]. Machine Learning，1996，24(2)：123–140.
[2]	AFRAEI S，SHAHRIAR K，MADANI S H. Developing intelligent classification models for rock burst prediction after recognizing significant predictor variables，Section 1：Literature review and data preprocessing procedure[J]. Tunnelling and Underground Space Technology，2019，83：324–353.
[4]	陈烨开. 基于能量耗散理论的岩石损伤本构关系研究[硕士学位论文][D]. 淮南：安徽理工大学，2020.(CHEN Yekai. Research on constitutive relationship of rock damage based on energy dissipation theory[M. S. Thesis][D]. Huainan：Anhui University of Science and Technology，2020.(in Chinese))
[5]	BIENIAWSKI Z T，DENKHAUS H G，VOGLER U W. Failure of fractured rock[J]. International Journal of Rock Mechanics and Mining Sciences and Geomechanics，1969，3(6)：323–341.
[11]	李玉生. 矿山冲击名词探讨–兼评“冲击地压”[J]. 煤炭学报，1982，3(2)：89–95.(LI Yusheng. Discussion on the terminology of mine shock-comment on“burst pressure”[J]. Journal of China Coal Society，1982，3(2)：89–95.(in Chinese))
[1]	杨剑锋，乔佩蕊，李永梅，等. 机器学习分类问题及算法研究综述[J]. 统计与决策，2019，35(6)：36–40.(YANG Jianfeng，QIAO Peirui，LI Yongmei，et al. A review of machine learning classification problems and algorithms[J]. Statistics and Decision，2019，35(6)：36–40.(in Chinese))
[3]	贾义鹏. 岩爆预测方法与理论模型研究[博士学位论文][D]. 杭州：浙江大学，2014.(JIA Yipeng. Research on rockburst prediction methods and theoretical models[Ph. D. Thesis][D]. Hangzhou：Zhejiang University，2014.(in Chinese))
[6]	于群. 深埋隧洞岩爆孕育过程及预警方法研究[博士学位论文][D]. 大连：大连理工大学，2016.(YU Qun. Research on the process of rockburst in deep-buried tunnels and early warning methods[[Ph. D. Thesis][D]. Dalian：Dalian University of Technology，2016.(in Chinese))
[7]	章梦涛，徐曾和，潘一山，等. 冲击地压和突出的统一失稳理论[J]. 煤炭学报，1991，(4)：48–53.(ZHANG Mengtao，XU Zenghe，PAN Yishan，et al. Unified instability theory of rock burst and outburst[J]. Journal of China Coal Society，1991，(4)：48–53.(in Chinese))
[9]	乔趁. 基于突变理论的深埋隧道岩爆危险性预测研究[硕士学位论文][D]. 邯郸：河北工程大学，2018.(QIAO Chen. Research on the prediction of rockburst hazard in deep-buried tunnels based on catastrophe theory[M. S. Thesis][D]. Handan：Hebei University of Engineering，2018.(in Chinese))
[12]	李鹏翔，陈炳瑞，周扬一，等. 硬岩岩爆预测预警研究进展[J]. 煤炭学报，2019，44(增2)：447–465.(LI Pengxiang，CHEN Binrui，ZHOU Yangyi，et al. Research progress in hard rock burst prediction and early warning[J]. Journal of China Coal Society，2019，44(Supp.2)：447–465.(in Chinese))
[17]	张乐文，张德永，李术才，等. 基于粗糙集理论的遗传–RBF神经网络在岩爆预测中的应用[J]. 岩土力学，2012，33(增1)：270–276.(ZHANG Lewen，ZHANG Deyong，LI Shucai，et al. Application of genetic-RBF neural network based on rough set theory in rockburst prediction[J]. Rock and Soil Mechanics，2012，33(Supp.1)：270–276.(in Chinese))
[23]	LIANG W Z，DAI B，ZHAO G Y，et al. A scientometric review on rockburst in hard rock：two decades of review from 2000 to 2019[J]. Geofluids，2020，2020：1–17.
[8]	胡峰. 基于断裂力学和损伤理论的裂隙岩体损伤机制研究[硕士学位论文][D]. 重庆：重庆大学，2015.(HU Feng. Research on damage mechanism of fractured rock mass based on fracture mechanics and damage theory[M. S. Thesis][D]. Chongqing：Chongqing University，2015.(in Chinese))
[13]	冯夏庭. 地下峒室岩爆预报的自适应模式识别方法[J]. 东北大学学报，1994，15(5)：471–475.(FENG Xiating. Adaptive pattern recognition method for underground cavern rockburst prediction[J]. Journal of Northeastern University，1994，15(5)：471–475.(in Chinese))
[15]	赵国彦，刘强，刘超. 岩爆烈度分级预测中的贝叶斯判别分析[J]. 金属矿山，2010，(5)：143–147.(ZHAO Guoyan，LIU Qiang，LIU Chao. Bayesian discriminant analysis in rockburst intensity classification prediction[J]. Metal Mine，2010，(5)：143–147.(in Chinese))
[16]	DONG L J，LI X B，PENG K. Prediction of rockburst classification using Random Forest[J]. Transactions of Nonferrous Metals Society of China，2013，23(2)：472–477.
[19]	郭庆清，刘磊磊，张绍和，等. 基于组合赋权法和聚类分析法的岩爆预测[J]. 长江科学院院报，2013，30(12)：54–59.(GUO Qinqing，LIU Leilei，ZHANG Shaohe，et al. Rockburst prediction based on combined weighting method and cluster analysis method[J]. Journal of Yangtze River Scientific Research Institute，2013，30(12)：54–59.(in Chinese))
[21]	YIN X，LIU Q S，PAN Y C，et al. Strength of stacking technique of ensemble learning in rockburst prediction with imbalanced data：comparison of eight single and ensemble models[J]. Natural Resources Research，2021，30(2)：1 795–1 815.
[24]	刘晓悦，季红瑜. 基于AdaBoost-BAS-SVM模型的岩爆预测研究[J]. 金属矿山，2021，45(10)：28–34.(LIU Xiaoyue，JI Hongyu. Research on rockburst prediction based on AdaBoost-BAS-SVM model[J]. Metal Mine，2021，45(10)：28–34.(in Chinese))
[18]	邱道宏，李术才，张乐文，等. 基于模型可靠性检查的QGA-SVM岩爆倾向性分类研究[J]. 应用基础与工程科学学报，2015，23(5)：981–991.(QIU Daohong，LI Shucai， ZHANG Lewen，et al. Research on QGA-SVM rockburst tendency classification based on model reliability check[J]. Journal of Applied Basic Science and Engineering，2015，23(5)：981–991.(in Chinese))
[20]	ZHANG J F，WANG Y H，SUN Y T，et al. Strength of ensemble learning in multiclass classification of rockburst intensity[J]. International Journal for Numerical and Analytical Methods in Geomechanics，2020，44(13)：1 833–1 853.
[22]	WANG S M，ZHOU J，LI C Q，et al. Rockburst prediction in hard rock mines developing bagging and boosting tree-based ensemble techniques[J]. Journal of Central South University，2021，28(2)：527–542.
[28]	张淑清，段晓宁，张立国，等. Tsne降维可视化分析及飞蛾火焰优化ELM算法在电力负荷预测中应用[J]. 中国电机工程学报，2021，41(9)：3 120–3 130.(ZHANG Shuqing，DUAN Xiaoning，ZHANG Liguo，et al. Application of Tsne dimensionality reduction visualization analysis and moth flame optimization ELM algorithm in power load forecasting[J]. Proceedings of the Chinese Society of Electrical Engineering，2021，41(9)：3 120–3 130.(in Chinese))
[31]	SCHAPIRE R E，FREUND P. Boosting the margin: A new explanation for the effectiveness of voting methods[J]. Annals of Statistics，1998，26(5)：1 651–1 686.
[25]	谭文侃，叶义成，胡南燕，等. LOF与改进SMOTE算法组合的强烈岩爆预测[J]. 岩石力学与工程学报，2021，40(6)：1 186–1 194. (TAN Wenkan，YE Yicheng，HU Nanyan，et al. Strong rockburst prediction based on the combination of LOF and improved SMOTE algorithm[J]. Chinese Journal of Rock Mechanics and Engineering，2021，40(6)：1 186–1 194.(in Chinese))
[26]	田睿，孟海东，陈世江，等. 基于深度神经网络的岩爆烈度分级预测[J]. 煤炭学报，2020，45(增1)：191–201.(TIAN Rui，MENG Haidong，CHEN Shijiang，et al. Rockburst intensity classification prediction based on deep neural network[J]. Journal of China Coal Society，2020，45(Supp.1)：191–201.(in Chinese))
[27]	WU S C，WU Z G，ZHANG C X. Rock burst prediction probability model based on case analysis[J]. Tunnelling and Underground Space Technology incorporating Trenchless Technology Research，2019，93(C)：1–15.
[30]	KEARNS M，VALIANT L. Cryptographic limitations on learning Boolean formulae and finite automata[J]. Journal of the ACM，1994，41(1)：67–95.
[29]	刘斌. 基于自注意力机制的文本分类研究[硕士学位论文][D]. 北京：中国地质大学，2020.(LIU Bin. Research on text classification based on self-attention mechanism[M. S. Thesis][D]. Beijing：China University of Geosciences，2020.(in Chinese))
[33]	谢学斌，李德玄，孔令燕，等. 基于CRITIC-XGB算法的岩爆倾向等级预测模型[J]. 岩石力学与工程学报，2020，39(10)：1 975–1 982. (XIE Xuebin，LI Dexuan，KONG Linyan，et al. Rockburst tendency grade prediction model based on CRITIC-XGB algorithm[J]. Chinese Journal of Rock Mechanics and Engineering，2020，39(10)：1 975–1 982.(in Chinese))