| 摘要 |
A corpus of documents is identified, such as a large corpus of web documents. A quality score is applied to each, and at least some of the documents in the corpus of documents are identified based on their respective quality scores. At least one query characteristic, for instance, the language of a query, associated with a plurality of search queries is identified. A subset of documents in the corpus of documents is identified that satisfy the at least one query characteristic. An index is built that includes the identified at least some documents and the identified subset of documents. |
