Scholarlink

<p>In the process of applying decision tree models to small sample training data for continuous casting, the excessive and deep branches make it difficult to effectively identify steel leakage. To solve this problem, a decision tree model optimized by genetic algorithm based on one vs rest (OvR) is established for predicting steel leakage. By using genetic algorithm to prune and configure the decision tree model, strengthening the supervision of the splitting and growth process of the model, and the accuracy of the model is improved. And on-site production data to test and compare the OvR-GA-CART model with traditional decision trees, GA-CART, and GA-ACO-BP models have been used. Tests have shown that the CART decision tree model combines the advantages of genetic algorithm and OvR method, achieving higher accuracy and generalization. It can achieve a 100% reporting rate for steel leakage and an overall accuracy of 98.39%, demonstrating good predictive performance and application prospects.</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

Application of OvR-GA-CART Decision Tree Algorithm in Predicting Steel Leakage in thin Slab Continuous Casting

  • 期刊论文

  • 发表日期：

    2025年2月26日

查看全文
  • Zhang Benguo,
  • Yu Haochen,
  • Zhao jie,
  • Zhang Ruizhong

摘要

In the process of applying decision tree models to small sample training data for continuous casting, the excessive and deep branches make it difficult to effectively identify steel leakage. To solve this problem, a decision tree model optimized by genetic algorithm based on one vs rest (OvR) is established for predicting steel leakage. By using genetic algorithm to prune and configure the decision tree model, strengthening the supervision of the splitting and growth process of the model, and the accuracy of the model is improved. And on-site production data to test and compare the OvR-GA-CART model with traditional decision trees, GA-CART, and GA-ACO-BP models have been used. Tests have shown that the CART decision tree model combines the advantages of genetic algorithm and OvR method, achieving higher accuracy and generalization. It can achieve a 100% reporting rate for steel leakage and an overall accuracy of 98.39%, demonstrating good predictive performance and application prospects.