摘要
支持向量机是一种基于统计学习理论的机器学习方法,具有很好的回归和预测性能。本文提出了一种基于支持向量机回归的结构损伤识别方法,采用柔度作为支持向量机的输入向量,并以两跨连续梁为例进行仿真计算。结果表明,本文方法可以较好的从单点损伤情况预测出两点损伤情况的损伤位置和损伤程度。
Support Vector Machine（SVM） is a machine learning algorithm based on statistical learn- ing theory with the good regression and generalization ability. A method for structural damage identifica- tion based on SVM regression is proposed in this paper, and the flexibility matrix was used as the SVM input vectors in an emulational example of two - span beam. The results show that the proposed method performs a good ability of predicting the location and maanitude of damages.
出处
《低温建筑技术》
2013年第9期45-47,共3页
Low Temperature Architecture Technology
关键词
损伤识别
支持向量机
柔度矩阵
damage identification
support vector machine
flexibility matrix
