Abstract: Ontology mapping is a focus research field today，even the computation of concept similarity is the key part to it.Most algorithms combine multiple strategies，however these algorithms have some problems such as computing largely，setting weight hardly etc.This paper proposes a modified algorithm－PCASim.It reduces computation by computing similarity of concept name and modifies weight by PCA.The experiment shows that the algorithm is viable.

Key words: WordNet, similarity, Principal Component Analysis（PCA）

摘要： 本体映射是目前的热点问题，而概念相似度计算则是它的关键部分。目前的方法基本上都是基于多策略的综合方法，而综合方法存在计算量大、权值难以确定等问题。提出了一种改进的综合算法－PCASim。该方法通过概念名称相似度计算减少计算量，利用主成分分析改进权值。实验表明该方法是切实可行的。

关键词: WordNet, 相似度, 主成分分析