Relevance vector machine with tuning based on self-adaptive differential evolution approach for predictive modelling of a chemical process

Acosta, Simone Massulini; Amoroso, Anderson Levati; Anna, Angelo Marcio Oliveira Sant; Canciglieri, Osíris

doi:10.1016/j.apm.2021.01.057

Cited by 22 publications

(8 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[24,25] Some research validated better performance of the RVM method than artificial neural network, ridge regression, random forest, K-nearest neighbors, and multiple linear regression models on the prediction of end-point phosphorus content, carbon content, and end-point temperature of molten steel during the steelmaking process. [5,26,27] The proposed FRVM method is the improvement of the RVM method with the FDA theory. Detailed calculation procedures of the FRVM methodology are expounded in this section and the evaluation criteria for model performance are given.…”

Section: Algorithm Of Functional Relevance Vector Machinementioning

confidence: 99%

“…For example, F. He et al established a model based on the principal component analysis (PCA) and back propagation neural network to predict the end-point phosphorus content of molten steel [3] ; R. H. Zhang et al compared the prediction accuracy of five machine learning models and a metallurgical mechanism model on the end-point phosphorus content [4] ; S. M. Acosta et al applied the relevance vector machine model to predict the end-point phosphorus content. [5] The model prediction way has the advantage of evaluating the molten steel's phosphorus content in time for process control and guaranteeing the continuity of the production process, but its prediction accuracy is difficult to be improved due to the complex rephosphorization phenomenon in the steelmaking process. At the same time, the inaccurate composition of raw materials and auxiliary materials, experimental process parameter control, and unstable production state bring about large statistical uncertainty of the prediction result of phosphorus The ORCID identification number(s) for the author(s) of this article can be found under https://doi.org/10.1002/srin.202300351.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Dynamic Prediction with Statistical Uncertainty Evaluation of Phosphorus Content Based on Functional Relevance Vector Machine

Qian,

Chang,

Dong

et al. 2024

steel research int.

View full text Add to dashboard Cite

In the steelmaking process, inaccurate composition of raw materials, experimental process parameter control, and other factors bring about large statistical uncertainty to the phosphorus content prediction of molten steel. Meanwhile, lots of models mainly predict the end‐point phosphorus content, but the composition change during the whole production process is still in the gray box condition, where the exact refining state is unknown. These two problems increase the difficulty in process parameter optimization and control. In order to improve the phosphorus content control in the steelmaking process, this work proposes the functional relevance vector machine (FRVM) method. This method smooths the time series of process parameters to continuous functions by functional data analysis to predict the phosphorus content change during the whole production process, and estimates the probabilistic distributions of regression weights based on the Bayesian framework to directly output the probabilistic confidence interval of the phosphorus content. The real‐world steelmaking dataset validates that the mean relative prediction error of phosphorus content by the FRVM method is less than 18.04% during the whole production process and the end‐point hit rate is 86.46% when the prediction error is within ±0.005 wt%.

show abstract

Section: Algorithm Of Functional Relevance Vector Machinementioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Dynamic Prediction with Statistical Uncertainty Evaluation of Phosphorus Content Based on Functional Relevance Vector Machine

Qian,

Chang,

Dong

et al. 2024

steel research int.

View full text Add to dashboard Cite

show abstract

“…In addition, while maintaining the properties of SVM, RVM gets around issues with estimating the penalty coefficient of SVM and the need for the kernel function to satisfy Mercer's theorem. It is easier to use and makes predictions more accurate [21]. In choosing the optimal RVM kernel function, as compared to the single kernel learning approach, the use of multiple kernel functions may prevent blind selection.…”

Section: Construction Of Hkrvmmentioning

confidence: 99%

Research on Adaptive TSSA-HKRVM Model for Regression Prediction of Crane Load Spectrum

Qian¹,

Qin²,

Huang³

et al. 2023

Computer Modeling in Engineering &Amp; Sciences

View full text Add to dashboard Cite

For the randomness of crane working load leading to the decrease of load spectrum prediction accuracy with time, an adaptive TSSA-HKRVM model for crane load spectrum regression prediction is proposed. The heterogeneous kernel relevance vector machine model (HKRVM) with comprehensive expression ability is established using the complementary advantages of various kernel functions. The combination strategy consisting of refraction reverse learning, golden sine, and Cauchy mutation + logistic chaotic perturbation is introduced to form a multi-strategy improved sparrow algorithm (TSSA), thus optimizing the relevant parameters of HKRVM. The adaptive updating mechanism of the heterogeneous kernel RVM model under the multi-strategy improved sparrow algorithm (TSSA-HKMRVM) is defined by the sliding window design theory. Based on the sample data of the measured load spectrum, the trained adaptive TSSA-HKRVM model is employed to complete the prediction of the crane equivalent load spectrum. Applying this method to QD20/10 t × 43 m × 12 m general bridge crane, the results show that: compared with other prediction models, although the complexity of the adaptive TSSA-HKRVM model is relatively high, the prediction accuracy of the load spectrum under long periods has been effectively improved, and the completeness of the load information during the whole life cycle is relatively higher, with better applicability.

show abstract

“…Thus, different functions can be used and compared to other activation functions (Du and Swamy, 2016). Acosta et al. (2021) used an ANN with Levenberg–Marquardt scheme for improving the energetic efficiency and the influence of phosphorus levels on refining process.…”

Section: Machine Learning Modelsmentioning

confidence: 99%

Machine learning-based control charts for monitoring fraction nonconforming product in smart manufacturing

Acosta

Sant’Anna

2022

IJQRM

Self Cite

View full text Add to dashboard Cite

PurposeProcess monitoring is a way to manage the quality characteristics of products in manufacturing processes. Several process monitoring based on machine learning algorithms have been proposed in the literature and have gained the attention of many researchers. In this paper, the authors developed machine learning-based control charts for monitoring fraction non-conforming products in smart manufacturing. This study proposed a relevance vector machine using Bayesian sparse kernel optimized by differential evolution algorithm for efficient monitoring in manufacturing.Design/methodology/approachA new approach was carried out about data analysis, modelling and monitoring in the manufacturing industry. This study developed a relevance vector machine using Bayesian sparse kernel technique to improve the support vector machine used to both regression and classification problems. The authors compared the performance of proposed relevance vector machine with other machine learning algorithms, such as support vector machine, artificial neural network and beta regression model. The proposed approach was evaluated by different shift scenarios of average run length using Monte Carlo simulation.FindingsThe authors analyse a real case study in a manufacturing company, based on best machine learning algorithms. The results indicate that proposed relevance vector machine-based process monitoring are excellent quality tools for monitoring defective products in manufacturing process. A comparative analysis with four machine learning models is used to evaluate the performance of the proposed approach. The relevance vector machine has slightly better performance than support vector machine, artificial neural network and beta models.Originality/valueThis research is different from the others by providing approaches for monitoring defective products. Machine learning-based control charts are used to monitor product failures in smart manufacturing process. Besides, the key contribution of this study is to develop different models for fault detection and to identify any change point in the manufacturing process. Moreover, the authors’ research indicates that machine learning models are adequate tools for the modelling and monitoring of the fraction non-conforming product in the industrial process.

show abstract

Relevance vector machine with tuning based on self-adaptive differential evolution approach for predictive modelling of a chemical process

Cited by 22 publications

References 54 publications

Dynamic Prediction with Statistical Uncertainty Evaluation of Phosphorus Content Based on Functional Relevance Vector Machine

Dynamic Prediction with Statistical Uncertainty Evaluation of Phosphorus Content Based on Functional Relevance Vector Machine

Research on Adaptive TSSA-HKRVM Model for Regression Prediction of Crane Load Spectrum

Machine learning-based control charts for monitoring fraction nonconforming product in smart manufacturing

Contact Info

Product

Resources

About