摘要
为了提高制造资源和服务的发现准确性,提出了一种基于互联网的制造网络节点发现方法.制造主题节点内容和结构存在相似的特点,从网络节点的文本中抽取反映节点主题的结构和内容特征词,采用不同的加权方法表示为内容和结构特征信息,建立描述节点主题的混合向量空间模型.在此模型的基础上,通过类中心距离法来分析和识别节点的主题,并利用网络搜索、信息处理等技术构建了制造节点发现的实验原型系统.实验结果表明,混合向量空间模型适合描述制造主题的节点,原型系统对制造节点的发现准确性较高.
The approach of Manufacturing node discovery was developed to improve the accuracy of manufacturing resource searching from the Web. The structures and contents of a node reflect its topic synthetically. A hybrid vector space model （HVSM） to denote nodal topic was proposed accordingly. This model identified the topics of different nodes through similarity measure. Its feature vector elements were composed of the feature information about the node structure and content, and extracted from relevant web pages. The prototype system based on HVSM model was built by web search and information processing techniques. Experimental results showed that this system has good performance in domain-specific node recognition. Furthermore, HVSM model is suitable for node topic description and easy to implement in actual applications.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2006年第5期738-742,909,共6页
Journal of Zhejiang University：Engineering Science
基金
国家自然科学基金资助项目(60374057
50575204)
国家"863"高科技研究发展计划资助项目(2005AA411910)
关键词
制造网络
向量空间模型
相似性
分类
manufacturing networks
vector space model, similarity
classification
