摘要 |
A document management system indexes categories for efficient retrieval based on a category sequence. The category sequence is generated such that for any first, second and third categories appearing in the sequence in ascending order, a similarity distance between the first and second categories is less than or equal to a similarity distance between the first and third categories, and a similarity distance between the second and third categories is also less than or equal to the similarity distance between the first and third categories. A category index implemented in this manner significantly reduces the number of similarity distance computations that are performed when searching for categories and documents that are most relevant to content presented on a web page.
|