检索中关键词对相关性及其算法

TERM-PAIR CORRELATION IN RETRIEVAL AND ITS ALGORITHM

文章在表示文献向量的关键词向量组相关的假设下，提出了非正交基的向最空间检索模型的概念，并导出计算关键词对相关性的相关矩阵和相应匹配函数的算法。这可作为索引关键词的自动分类和改善各种检索策略中关键词相关性的基础。

On the assumption that there is correlation among term vectors for rep- resenting document vectors, this paper proposes the concept of non-orthogohal vector space model for representing document vectors, thereby, deducing the correlation matrix for calculating the term-pair correlationship and the corresponding matching function algorithm which can be used for the automatic classification of indexing terms and serves as the basis tot improving the index term correlationship among various strategies.