A framework based on heterogeneous ensemble models for liquid steel temperature prediction in LF refining process
Section snippets
Code metadata
Permanent link to reproducible Capsule: https://doi.org/10.24433/CO.3289734.v1.
Smelting process description and data collection
Before continuous casting, the compact steelmaking process mainly includes Consteel electric arc furnacedeoxidation and alloyingLF refining. The Consteel electric arc furnace is in the early stage of producing liquid steel, while the LF refining is to adjust the composition and temperature of liquid steel to improve product quality and continuous casting effect. The deoxidation and alloying is an important stage linking the Consteel electric arc furnace and the LF refining, in which the
Review of tree-based ensemble models
Typical representatives of tree-based ensemble models, including Random Forests [16], Extremely Randomized Trees (Extra Trees) [17], eXtreme Gradient Boosting (XGBoost) [18], Gradient Boosting Decision Tree (GBDT) [19], and AdaBoost-CART [20], are adopted as main candidate models for the heterogeneous ensemble model. Tree-based ensemble models include two main frameworks, Bagging and Boosting [21], [22], [23], [24], which are better strategies to trade off the fitting and generalization ability
Construction of heterogeneous ensemble models
The heterogeneous ensemble model mainly includes two model structures, Stacking and Majority Voting (MV), as shown in Fig. 3. The main difference between Stacking and MV is that in layer 2, Stacking requires a suitable meta-learner to integrate the results of layer 1 to form the final result, while MV only votes for the final result based on the results of layer 1.
To facilitate the construction of the heterogeneous ensemble model, Recursive Feature Increase (RFI) is proposed in this paper. The
Hyper-parameter optimization of heterogeneous ensemble models
In the above section, the RFI is employed to construct the heterogeneous ensemble model. In this section, we will optimize the hyper-parameters of the heterogeneous ensemble model to further improve the predictive performance.
Conclusion
In this study, a novel framework based on heterogeneous ensemble models is proposed for the liquid steel temperature prediction in the LF refining process. In this process, various machine learning models, including single models and tree-based ensemble models, are selected as candidate models for the heterogeneous ensemble model, and meanwhile, it is found that tree-based ensemble models possess better predictive performance for the liquid steel temperature prediction in the LF refining
CRediT authorship contribution statement
Chao Chen: Conceptualization, Methodology, Data curation, Project administration, Investigation, Validation, Writing – review & editing, Software. Nan Wang: Writing – review & editing, Project administration, Funding acquisition, Supervision. Min Chen: Writing – review & editing, Project administration, Funding acquisition, Supervision. Xumei Yan: Writing – review & editing, Data curation.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
The authors gratefully acknowledge the National Natural Science Foundation of China. [Grant numbers: 52074077, 52174301, and 51974080] and the Fundamental Research Funds for the Central Universities supported by the Chinese Education Ministry [Grant number N2125018].
References (49)
- et al.
Ladle furnace liquid steel temperature prediction model based on optimally pruned Bagging
J. Iron Steel Res. Int.
(2012) - et al.
Molten steel temperature prediction model based on bootstrap feature subsets ensemble regression trees
Knowl.-Based Syst.
(2016) - et al.
Featured temporal segmentation method and AdaBoost-BP detector for internal leakage evaluation of a hydraulic cylinder
Measurement
(2018) - et al.
A novel surrogate model for channel geometry optimization of PEM fuel cell based on Bagging-SVM ensemble regression
Int. J. Hydrog. Energy
(2022) - et al.
Wind speed prediction using measurements from neighboring locations and combining the extreme learning machine and the AdaBoost algorithm
Energy Rep.
(2022) - et al.
Interpretable boosting tree ensemble method for multisource building fire loss prediction
Reliab. Eng. Syst. Saf.
(2022) - et al.
An improved stacking ensemble learning-based sensor fault detection method for building energy systems using fault-discrimination information
J. Build. Eng.
(2021) - et al.
Multi-step forecasting for big data time series based on ensemble learning
Knowl.- Based Syst.
(2019) - et al.
Novel application of multi-model ensemble learning for fault diagnosis in refrigeration systems
Appl. Therm. Eng.
(2020) - et al.
A novel heterogeneous ensemble credit scoring model based on bstacking approach
Expert Syst. Appl.
(2018)
Based on multi-algorithm hybrid method to predict the slope safety factor-stacking ensemble learning with Bayesian optimization
J. Comput. Sci-Neth.
A stacking-based ensemble learning method for earthquake casualty prediction
Appl. Soft Comput.
A decision-theoretic generalization of on-line learning and an application to boosting
J. Comput. System Sci.
A practical tutorial on bagging and boosting based ensembles for machine learning: Algorithms, software tools, performance study, practical perspectives and opportunities
Inform. Fusion
Relevance vector machine with tuning based on self-adaptive differential evolution approach for predictive modelling of a chemical process
Appl. Math. Model.
Practical selection of SVM parameters and noise estimation for SVM regression
Neural Netw.
Using support vector regression and k-nearest neighbors for short-term traffic flow prediction based on maximal information coefficient
Inform. Sciences
Prediction model of rock mass class using classification and regression tree integrated AdaBoost algorithm based on TBM driving data
Tunn. Undergr. Space Technol.
Fair-AdaBoost: Extending AdaBoost method to achieve fair classification
Expert Syst. Appl.
Novel hybrid ensemble credit scoring model with stacking-based noise detection and weight assignment
Expert Syst. Appl.
Fast stepwise regression based on multidimensional indexes
Inform. Sci.
Cylinder drag minimization through wall actuation: A Bayesian optimization approach
Comput. Fluids
Fast hyperparameter tuning using Bayesian optimization with directional derivatives
Knowl.-Based Syst.
On the performance of artificial bee colony (ABC) algorithm
Appl. Soft Comput.
Cited by (3)
A new hybrid credit scoring ensemble model with feature enhancement and soft voting weight optimization
2024, Expert Systems with ApplicationsGrid search with a weighted error function: Hyper-parameter optimization for financial time series forecasting
2024, Applied Soft ComputingResearch on feature modeling method for complex industrial process and its application
2023, Kongzhi yu Juece/Control and Decision
The code (and data) in this article has been certified as Reproducible by Code Ocean: (https://codeocean.com/). More information on the Reproducibility Badge Initiative is available at https://www.elsevier.com/physical-sciences-and-engineering/computer-science/journals.