基于支持向量机的混凝土结构中钢筋腐蚀的判别

提出了一种基于多因素的混凝土中钢筋腐蚀判别方法.该方法综合考虑影响钢筋腐蚀的多个因素的共同作用,克服了以往单因素判别钢筋腐蚀不够客观的弊端.本文首先通过相关分析比较各个影响因素的相关性,然后通过变量聚类对多个影响因素分类,选出典型变量建立基于支持向量机的钢筋腐蚀状态分类器,最后在实际工程中检验了其判断的准确性,并对比了以往曾有学者根据Fisher准则建立的分类器的判别结果.结果表明,基于支持向量机的分类器分类准确率优于基于Fisher准则的分类器.

Evaluating Corrosion Severity of Reinforcing Steel in Concrete Based on Support Vector Machine

A method of evaluating corrosion in reinforced concrete is introduced,it takes important factors that have close relationship with corrosion severity into account.First,it compares the correlations between factors.Second,it classifies the impact factors by cluster analysis and picks out the most representative factors.And then,it establishes the classification model of evaluating corrosion severity based on support vector machine.At last,we compared the performances of classifier based on SVM with the classifier based on Fisher's Principle.The results show that the classification accuracy rate of classifier based on SVM is prior to that based on Fisher's Principle.