Compositional SVM Model Based on Bayesian Classifier and Linear Discriminant Analysis

摘要：

Data Mining is the most important aspect in building an object model for a complex industrial process. The existing data mining classification methods are mostly based on the input process object information while neglecting the output information, which leads to incomplete data information and inaccuracy of the object model. This article proposes a method by introducing Linear Discriminant Analysis to Bayesian classifier to establish the model. The classifier is built by using Naive Bayesian algorithm, in which the object elements includes input and output information. An improved LDA algorithm is used to transform the differentiated classes of data to reduce the interference information of the adjacent classes. An industrial applied instance shows that the model has higher estimation accuracy and better tracking capability.

关键词： Bayesian classifier Linear Discriminant Analysis Compositional model SVM

作者: Yujun Deng Huizhong Yang

作者单位: School of Communication and Engineering, Jiangnan University, Wuxi 214122, JiangSu,China School of Communication and Engineering, Jiangnan University, Wuxi 214122, JiangSu,China

会议类型: 国际会议

会议名称: The First World Congress on Global Optimization in Engineering & Science(第一届工程与科学全局优化国际会议 WCGO2009)

会议地点: 长沙

会议语种:英文

页码: 413-420

在线出版日期: 2009-06-01（万方平台首次上网日期，不代表论文的发表时间）