工程类几何模型分类与聚类的现状与趋势  

Classification and clustering of geometry models in engineering:state-of-the-art review and future trends

作　　者：莫蓉[1] 石源[1] 常智勇[1] 陈泽峰[1] 张欣[1] 

机构地区：[1]西北工业大学现代设计与集成制造技术教育部重点实验室,陕西西安710072

出　　处：《计算机集成制造系统》2010年第3期449-461,共13页Computer Integrated Manufacturing Systems

基　　金：国家自然科学基金资助项目(50805122);国家863/CI MS主题资助项目(2007AA04Z184)~~

摘　　要：总结了目前工程类几何模型分类与聚类问题的研究进展。首先分析了基于k最近邻方法和基于支持向量机的模型分类技术,回顾了传统聚类技术和人工神经网络在模型聚类方面的应用现状;然后探讨了聚类和降维技术在模型分类与聚类的过程可视化和结果可视化方面的应用。最后,通过对已有研究成果的比较分析,预测了工程类几何模型的分类与聚类的研究方向。Targeting on the issues about classification and clustering of geometry models in engineering field,current research development was summarized.Firstly,model classification based on k-Nearest Neighbor approach and Support Vector Machine were analyzed.Then the review of model clustering technology relevant to traditional Clustering Analysis and Artificial Neural Networks was presented.Furthermore,clustering and dimension reduction technologies,which were used to visualize classification/clustering process and results,were discussed.Through a comparative analysis of existing research results,the future development trends of engineering geometry model classification and clustering were predicted finally.

关 键 词：几何模型 分类 聚类 可视化 数据挖掘 

分 类 号：TP391.71[自动化与计算机技术—计算机应用技术]