Scholarlink

<p>Given the difficulty in determining the parameters of the compressive strength prediction model of self-compacting concrete and the low prediction accuracy, this study focuses on the applicability of the relevance vector machine (RVM) model constructed using various optimization techniques in predicting the strength of self-compacting concrete. The principal component analysis (PCA) is first used to reduce the dimension of the influencing factors. Then, the particle swarm optimization algorithm (PSO) is introduced into the RVM to establish a PCA–PSO–RVM collaborative optimization model, which is compared with the traditional regression model through various statistical indicators and error analysis. The results show that the collaborative optimization model prediction based on PCA–PSO–RVM performs outstandingly in all performance indicators. In the test set, the R<sup>2</sup> of the collaborative optimization model is 0.978, MAE is 0.123, MSE is 0.021, and RMSE is 0.150. The evaluation of quantitative indicators verifies that the collaborative optimization model is feasible and advanced in predicting the strength of self-compacting concrete. This study also provides a reference for the research on durability, rheological properties, and other material predictions of self-compacting concrete.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Strength Prediction of Self-Compacting Concrete Using Improved RVM Machine Learning Method

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月24日

查看全文
  • Yan Zhang,
  • Yulong Ye,
  • Junfeng Wang,
  • Beichang Tang,
  • Feng Fu

摘要

Given the difficulty in determining the parameters of the compressive strength prediction model of self-compacting concrete and the low prediction accuracy, this study focuses on the applicability of the relevance vector machine (RVM) model constructed using various optimization techniques in predicting the strength of self-compacting concrete. The principal component analysis (PCA) is first used to reduce the dimension of the influencing factors. Then, the particle swarm optimization algorithm (PSO) is introduced into the RVM to establish a PCA–PSO–RVM collaborative optimization model, which is compared with the traditional regression model through various statistical indicators and error analysis. The results show that the collaborative optimization model prediction based on PCA–PSO–RVM performs outstandingly in all performance indicators. In the test set, the R2 of the collaborative optimization model is 0.978, MAE is 0.123, MSE is 0.021, and RMSE is 0.150. The evaluation of quantitative indicators verifies that the collaborative optimization model is feasible and advanced in predicting the strength of self-compacting concrete. This study also provides a reference for the research on durability, rheological properties, and other material predictions of self-compacting concrete.