摘要
为了确定专利文献的相似性,帮助企业进行专利申请、保护和利用,提出基于向量空间模型(VSM)和专利文献特征的相似专利确定方法.依据专利文献的信息特征构建专利模型树,定义了专利模型树和专利模型树的节点.通过分析专利模型树的节点属性值,采用基于向量空间模型的文本分类技术,以专利名称和专利摘要的加权相似度作为专利文献分类的依据,对专利文献进行分类,然后在类内根据专利文献特征的相似性确定相似专利,并根据企业的实际应用需求,分析专利文献要素权重确定的几种方法.应用示例验证了该方法能够有效地进行专利分类和相似专利检索.
A method to discover the similarity of patent documents was proposed in order to help enterprises in patent application, protection and utilization. A patent model tree was built based on the characteristics of patent documents. The patent model tree and its nodes were defined. Through analyzing the nodes＇ attribute values, patent documents were categorized by using the vector space model（VSM） based text categorization technology and the weighted similarities of patent name and patent abstract. According to the categorization, similar patents were discovered by the weighted similarities of patent characteristics in the same category. Several ways to identify the weight of patent characteristics were discussed according to the actual needs in enterprise application. A case study showed that the method can be used in patent categorization and similar patent search.
出处
《浙江大学学报（工学版）》
EI
CAS
CSCD
北大核心
2009年第10期1848-1852,1869,共6页
Journal of Zhejiang University：Engineering Science
基金
国家“十一五”科技支撑计划资助项目(2006BAF01A02)
国家“863”高技术研究发展计划资助项目(2007AA04Z101)
关键词
专利文献
专利检索
文本分类
向量空间模型
patent documents
patent retrieve
text categorization
vector space model（VSM）
