1. Faculty of Architecture, Civil and Transportation Engineering, Beijing University of Technology, Beijing 100124, China 2. School of Civil Engineering, Southeast University, Nanjing 211189, China
To reliably and accurately predict the axial bearing capacity of concrete-filled steel tube (CFST) columns, a prediction model of CFST column axial bearing capacity with ensemble machine learning was developed and explained. The quality of the CFST column database was evaluated using the Mahalanobis distance, the prediction model of CFST column axial bearing capacity was established by the extreme gradient boosting (XGBoost) algorithm, and the optimal hyperparameter combination of the model was found using the K-Fold cross-validation (K-Fold CV) and the tree-structured Parzen estimator (TPE) algorithms. The predicted values of the optimized XGBoost model were compared with the calculated values of the existing methods and the unoptimized XGBoost model using different evaluation metrics. The Shapley additive explanations (SHAP) approach was used to produce both global and local explanations for the predictions of XGBoost model. Results show that, after hyperparameter tuning, the XGBoost model’s performance surpasses performance of relevant standards and empirical formulas, and the SHAP approach can effectively explain the XGBoost model’s output.
Xi-ze CHEN,Jun-feng JIA,Yu-lei BAI,Tong GUO,Xiu-li DU. Prediction model of axial bearing capacity of concrete-filled steel tube columns based on XGBoost-SHAP. Journal of ZheJiang University (Engineering Science), 2023, 57(6): 1061-1070.
Tab.1Statistical information on parameters of concrete-filled steel tube column database
Fig.1Parameter frequency distribution diagram for concrete-filled steel tube column database
Fig.2Pearson correlation coefficient matrix for input and output parameters
Fig.3Mahalanobis distance for sample of concrete-filled steel tube column database
Fig.4Schematic description of K-Fold CV
Fig.5Searching for optimal hyperparameter ranges of model by 10-Fold CV
Fig.6Searching for optimal hyperparameter combinations of model by tree-structured Parzen estimator
Fig.7Demonstration of XGBoost model performance after hyperparameter optimisation
Fig.8Comparison between true and predicted values for nine methods at test dataset
模型
R2
RMSE
MAE
MAPE
文献[43]
0.854
315.651
231.706
27.047
文献[44]
0.768
397.773
323.111
38.605
文献[45]
0.785
382.387
320.920
36.873
文献[46]
0.857
382.387
320.920
36.873
文献[47]
0.854
315.720
244.594
36.944
文献[48]
0.844
325.718
251.056
37.057
文献[11]
0.920
233.976
195.502
24.078
文献[13]
0.922
230.704
176.394
18.555
本研究
0.963
157.958
93.989
10.388
Tab.2Performance evaluation metrics for nine models
Fig.9Changes in performance evaluation metrics before and after model optimization
Fig.10SHAP summary plot of six input parameters
Fig.11SHAP feature dependence plots of three input parameters
Fig.12Graph of SHAP prediction interpretation for single sample
[1]
HOU C, ZHOU X G Strength prediction of circular CFST columns through advanced machine learning methods[J]. Journal of Building Engineering, 2022, 51: 104289
doi: 10.1016/j.jobe.2022.104289
[2]
韩林海, 牟廷敏, 王法承, 等 钢管混凝土混合结构设计原理及其在桥梁工程中的应用[J]. 土木工程学报, 2020, 53 (5): 1- 24 HAN Lin-hai, MU Ting-min, WANG Fa-cheng, et al Design theory of CFST (concrete-filled steel tubular) mixed structures and its applications in bridge engineering[J]. China Civil Engineering Journal, 2020, 53 (5): 1- 24
doi: 10.15951/j.tmgcxb.20200413.001
[3]
李学彬, 杨仁树, 高延法, 等 大断面软岩斜井高强度钢管混凝土支架支护技术[J]. 煤炭学报, 2013, 38 (10): 1742- 1748 LI Xue-bin, YANG Ren-shu, GAO Yan-fa, et al High-strength steel tubular confined concrete supports support technology for large section soft rock inclined shaft[J]. Journal of China Coal Society, 2013, 38 (10): 1742- 1748
doi: 10.13225/j.cnki.jccs.2013.10.018
[4]
YIN F, WANG R W, CAO W L, et al Seismic behavior of light weight CFST frame-fabricated single row reinforced shear wall structures[J]. Structures, 2021, 34: 4486- 4499
doi: 10.1016/j.istruc.2021.10.044
[5]
陈志华, 杜颜胜, 吴辽, 等 矩形钢管混凝土结构研究综述[J]. 建筑结构, 2015, 45 (16): 40- 46 CHEN Zhi-hua, DU Yan-sheng, WU Liao, et al Review of the research on rectangular concrete-filled steel tube structures[J]. Building Structure, 2015, 45 (16): 40- 46
doi: 10.19701/j.jzjg.2015.16.009
[6]
LIEW J Y R, XIONG M, XIONG D Design of concrete filled tubular beam-columns with high strength steel and concrete[J]. Structures, 2016, 8: 213- 226
doi: 10.1016/j.istruc.2016.05.005
[7]
骆涛涛, 李娜, 张涛, 等 圆钢管混凝土柱轴心受压承载力计算[J]. 武汉大学学报:工学版, 2014, 47 (6): 764- 768 LUO Tao-tao, LI Na, ZHAO Tao, et al Bearing capacity of circular concrete-filled steel tube columns under axial compression[J]. Engineering Journal of Wuhan University, 2014, 47 (6): 764- 768
[8]
VU Q V, TRUONG V H, THAI H T Machine learning-based prediction of CFST columns using gradient tree boosting algorithm[J]. Composite Structures, 2021, 259: 113505
doi: 10.1016/j.compstruct.2020.113505
[9]
BOUKHATEM B, KENAI S, TAGNIT-HAMOU A, et al Application of new information technology on concrete: an overview[J]. Journal of Civil Engineering and Management, 2011, 17 (2): 248- 258
doi: 10.3846/13923730.2011.574343
[10]
NGUYEN H Q, LY H B, TRAN V Q, et al Optimization of artificial intelligence system by evolutionary algorithm for prediction of axial capacity of rectangular concrete filled steel tubes under compression[J]. Materials, 2020, 13 (5): 1205
doi: 10.3390/ma13051205
[11]
TRAN V L, THAI D K, KIM S E A new empirical formula for prediction of the axial compression capacity of CCFT columns[J]. Steel and Composite Structures, 2019, 33 (2): 181- 194
[12]
LIU X R, WU Y Q, ZHOU Y S Axial compression prediction and GUI design for CCFST column using machine learning and Shapley additive explanation[J]. Buildings, 2022, 12 (5): 698
doi: 10.3390/buildings12050698
[13]
JAVED M F, FAROOQ F, MEMON S A, et al New prediction model for the ultimate axial capacity of concrete-filled steel tubes: an evolutionary approach[J]. Crystals, 2020, 10 (9): 741
doi: 10.3390/cryst10090741
[14]
NASER M Z, THAI S, THAI H T Evaluating structural response of concrete-filled steel tubular columns through machine learning[J]. Journal of Building Engineering, 2021, 34: 101888
doi: 10.1016/j.jobe.2020.101888
[15]
LYU F, FAN X Y, DING F X, et al Prediction of the axial compressive strength of circular concrete-filled steel tube columns using sine cosine algorithm-support vector regression[J]. Composite Structures, 2021, 273: 114282
doi: 10.1016/j.compstruct.2021.114282
[16]
TAHMASSEBI A, MOTAMEDI M, ALAVI A H, et al An explainable prediction framework for engineering problems: case studies in reinforced concrete members modeling[J]. Engineering Computations, 2022, 39 (2): 609- 626
doi: 10.1108/EC-02-2021-0096
[17]
ZHOU Z H. Ensemble methods: foundations and algorithms [M]. New York: CRC press, 2012.
[18]
徐凯, 杨锐 基于QPSO-DBN集成学习的城轨列车定位研究[J]. 电子测量与仪器学报, 2022, 36 (9): 21- 28 XU Kai, YANG Rui Research on the positioning of urban rail train based on QPSO-DBN ensemble learning[J]. Journal of Electronic Measurement and Instrumentation, 2022, 36 (9): 21- 28
doi: 10.13382/j.jemi.B2104838
[19]
WEN L, HUGHES M Coastal wetland mapping using ensemble learning algorithms: a comparative study of bagging, boosting and stacking techniques[J]. Remote Sensing, 2020, 12 (10): 1683
doi: 10.3390/rs12101683
[20]
胡联粤, 岑健, 许文凯, 等 基于XGBoost算法的地下综合管廊安全状况评估方法[J]. 计算机应用, 2020, 40 (增1): 264- 268 HU Lian-yue, CEN Jian, XU Wen-kai, et al Safety situation evaluation method based on XGBoost algorithm for underground comprehensive pipe gallery[J]. Journal of Computer Applications, 2020, 40 (增1): 264- 268
[21]
LIM S, CHI S Xgboost application on bridge management systems for proactive damage estimation[J]. Advanced Engineering Informatics, 2019, 41: 100922
doi: 10.1016/j.aei.2019.100922
[22]
FENG D C, WANG W J, MANGALATHU S, et al Interpretable XGBoost-SHAP machine-learning model for shear strength prediction of squat RC walls[J]. Journal of Structural Engineering, 2021, 147 (11): 04021173
doi: 10.1061/(ASCE)ST.1943-541X.0003115
[23]
DARMATASIA, ARYMURTHY A M. Predicting the status of water pumps using data mining approach [C]// 2016 International Workshop on Big Data and Information Security (IWBIS). Jakarta: IEEE, 2016: 57-64.
[24]
冯德成, 吴刚 混凝土结构基本性能的可解释机器学习建模方法[J]. 建筑结构学报, 2022, 43 (4): 228- 238 FENG De-cheng, WU Gang Interpretable machine learning-based modeling approach for fundamental properties of concrete structures[J]. Journal of Building Structures, 2022, 43 (4): 228- 238
[25]
JIANG K, HAN Q, BAI Y, et al Data-driven ultimate conditions prediction and stress-strain model for FRP-confined concrete[J]. Composite Structures, 2020, 242: 112094
doi: 10.1016/j.compstruct.2020.112094
[26]
DUAN J, ASTERIS P G, NGUYEN H, et al A novel artificial intelligence technique to predict compressive strength of recycled aggregate concrete using ICA-XGBoost model[J]. Engineering with Computers, 2021, 37: 3329- 3346
doi: 10.1007/s00366-020-01003-0
[27]
CHANDRAMOULI P, JAYASEELAN R, PANDULU G, et al Estimating the axial compression capacity of concrete-filled double-skin tubular columns with metallic and non-metallic composite materials[J]. Materials, 2022, 15 (10): 3567
doi: 10.3390/ma15103567
[28]
NGO N T, PHAM T P T, LE H A, et al Axial strength prediction of steel tube confined concrete columns using a hybrid machine learning model[J]. Structures, 2022, 36: 765- 780
doi: 10.1016/j.istruc.2021.12.054
[29]
BARDHAN A, BISWAS R, KARDANI N, et al A novel integrated approach of augmented grey wolf optimizer and ANN for estimating axial load carrying-capacity of concrete-filled steel tube columns[J]. Construction and Building Materials, 2022, 337: 127454
doi: 10.1016/j.conbuildmat.2022.127454
[30]
DUNLOP P, SMITH S Estimating key characteristics of the concrete delivery and placement process using linear regression analysis[J]. Civil Engineering and Environmental Systems, 2003, 20 (4): 273- 290
doi: 10.1080/1028660031000091599
[31]
MAHALANOBIS P C. On the generalized distance in statistics [C]// Proceedings of the National Institute of Sciences. Caloutta: National Institute of Science of India, 1936, 2: 49-55.
[32]
QIAN W, GU C, ZHU C, et al. Short-term load forecasting based on multi-model fusion of CNN-LSTM-LGBM [C]// 2021 International Conference on Power System Technology. Haikou: IEEE, 2021: 934-939.
[33]
ZHU J, WANG Y Convolutional neural networks for predicting creep and shrinkage of concrete[J]. Construction and Building Materials, 2021, 306: 124868
doi: 10.1016/j.conbuildmat.2021.124868
[34]
LI W, YIN Y, QUAN X, et al Gene expression value prediction based on XGBoost algorithm[J]. Front Genet, 2019, 10: 1077
doi: 10.3389/fgene.2019.01077
[35]
CHEN T, GUESTRIN C. XGBoost: a scalable tree boosting system [C]// Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. [S.l.]: ACM, 2016: 785-794.
[36]
陈明华, 刘群英, 张家枢, 等 基于XGBoost的电力系统暂态稳定预测方法[J]. 电网技术, 2020, 44 (3): 1026- 1034 CHEN Ming-hua, LIU Qun-ying, ZHANG Jia-shu, et al XGBoost-based algorithm for post-fault transient stability status prediction[J]. Power System Technology, 2020, 44 (3): 1026- 1034
doi: 10.13335/j.1000-3673.pst.2018.1649
[37]
HUANG C L, CHEN M C, WANG C J Credit scoring with a data mining approach based on support vector machines[J]. Expert Systems with Applications, 2007, 33 (4): 847- 856
doi: 10.1016/j.eswa.2006.07.007
[38]
BERGSTRA J, BENGIO Y Random search for hyper-parameter optimization[J]. Journal of Machine Learning Research, 2012, 13: 281- 305
[39]
邓丽, 武金达, 李科学, 等. 基于TPE的SpaRC算法超参数优化方法[J]. 计算机科学. 2021, 48(2): 70-75. DENG Li, WU Jin-da, LI Ke-xue, et al. SpaRC algorithm hyperparameter optimization methodology based on TPE [J]. Computer Science, 2021, 48(2): 70-75.
[40]
BERGSTRA J, BÀRDENET R, BENGIO Y, et al. Algorithms for hyperparameter optimization [C]// Proceedings of the 24th Internation Conference on Neural Information Processing Systems. [S.l.]: Curran Associates Inc., 2011: 2546-2554.
[41]
DONG H, HE D, WANG F SMOTE-XGBoost using tree Parzen estimator optimization for copper flotation method classification[J]. Powder Technology, 2020, 375: 174- 181
doi: 10.1016/j.powtec.2020.07.065
[42]
BERGSTRA J, YAMINS D, COX D D. Making a science of model search: hyperparameter optimization in hundreds of dimensions for vision architectures [C]// Proceedings of the 30th International Conference on Machine Learning. Atlanta: [s.n.], 2013: 115-123.
British Standards Institution. Eurocode 4: design of composite steel and concrete structures: part 1-1: general rules and rules for buildings: BS EN 1994-1-1: 2004 [S]. London: British Standards Institution, 2004.
[45]
AIJ. Recommendations for design and construction of concrete-filled steel tubular structures [S]. [S.l.]: Architectural Institute of Japan, 2008.
[46]
ACI Committee 318. Building code requirements for structural concrete and commentary: ACI 318-08 [S]. Farmington Hills: ACI Committee Institute, 2008.
[47]
WANG Z B, TAO Z, HAN L H, et al Strength, stiffness and ductility of concrete-filled steel columns under axial compression[J]. Engineering Structures, 2017, 135: 209- 221
doi: 10.1016/j.engstruct.2016.12.049
[48]
GIAKOUMELIS G, LAM D Axial capacity of circular concrete-filled tube columns[J]. Journal of Constructional Steel Research, 2004, 60 (7): 1049- 1068
doi: 10.1016/j.jcsr.2003.10.001
[49]
XU X, LIU C, ZHAO Y, et al Short-term traffic flow prediction based on whale optimization algorithm optimized BiLSTM_Attention[J]. Concurrency and Computation: Practice and Experience, 2022, 34 (10): e6782
[50]
周挺, 杨军, 詹祥澎, 等 一种数据驱动的暂态电压稳定评估方法及其可解释性研究[J]. 电网技术, 2021, 45 (11): 4416- 4425 ZHOU Ting, YANG Jun, ZHAN Xiang-peng, et al Data-driven method and interpretability analysis for transient voltage stability assessment[J]. Power System Technology, 2021, 45 (11): 4416- 4425
doi: 10.13335/j.1000-3673.pst.2020.2170
[51]
LUNDBERG S M, LEE S I. A unified approach to interpreting model predictions [C]// Proceedings of the 31st International Conference on Neural Information Processing Systems. [S.l.]: Curran Associates Inc., 2017: 4768-4777.
[52]
NASIRI H, HOMAFAR A, CHELGANI S C Prediction of uniaxial compressive strength and modulus of elasticity for Travertine samples using an explainable artificial intelligence[J]. Results in Geophysical Sciences, 2021, 8: 100034
doi: 10.1016/j.ringps.2021.100034
