System and method for enhanced text matching
    4.
    发明授权
    System and method for enhanced text matching 有权
    用于增强文本匹配的系统和方法

    公开(公告)号:US07783660B2

    公开(公告)日:2010-08-24

    申请号:US11539040

    申请日:2006-10-05

    IPC分类号: G06F17/30

    摘要: The disclosure describes search systems and methods in which exact token searches, spelling suggestions, and split-token searches are used in conjunction to return search results to the user. Depending on the number and relevancy of results for the search query results from each of the steps the results are either merged or discarded into the final result set. The split-token search is adapted to generate two split-tokens from the token(s) of the search query in anticipation that the search token(s) is misspelled. As the location of the misspelling is unknown, the split-token search widens the scope of the results provided in response to the search. In an embodiment, the split-token search includes performing a prefix search for tokens matching a prefix split-token and a postfix search for tokens matching a postfix split-token. In an embodiment, the index is specially adapted to allow the postfix search to be performed more efficiently.

    摘要翻译: 本公开描述了搜索系统和方法,其中精确的令牌搜索,拼写建议和拆分令牌搜索被结合使用以将搜索结果返回给用户。 根据每个步骤的搜索查询结果的结果的数量和相关性,将结果合并或舍弃到最终结果集中。 分裂符号搜索适于从搜索查询的令牌生成两个拆分令牌,预期搜索令牌是拼写错误的。 由于拼写错误的位置未知,拆分令牌搜索扩大了响应搜索而提供的结果的范围。 在一个实施例中,分裂令牌搜索包括对匹配前缀拆分令牌的令牌和对匹配后缀拆分令牌的令牌的后缀搜索执行前缀搜索。 在一个实施例中,索引被特别地适于允许更有效地执行后缀搜索。

    SYSTEM AND METHOD FOR ENHANCED TEXT MATCHING
    5.
    发明申请
    SYSTEM AND METHOD FOR ENHANCED TEXT MATCHING 有权
    用于增强文本匹配的系统和方法

    公开(公告)号:US20080086488A1

    公开(公告)日:2008-04-10

    申请号:US11539040

    申请日:2006-10-05

    IPC分类号: G06F7/00

    摘要: The disclosure describes search systems and methods in which exact token searches, spelling suggestions, and split-token searches are used in conjunction to return search results to the user. Depending on the number and relevancy of results for the search query results from each of the steps the results are either merged or discarded into the final result set. The split-token search is adapted to generate two split-tokens from the token(s) of the search query in anticipation that the search token(s) is misspelled. As the location of the misspelling is unknown, the split-token search widens the scope of the results provided in response to the search. In an embodiment, the split-token search includes performing a prefix search for tokens matching a prefix split-token and a postfix search for tokens matching a postfix split-token. In an embodiment, the index is specially adapted to allow the postfix search to be performed more efficiently.

    摘要翻译: 本公开描述了搜索系统和方法,其中精确的令牌搜索,拼写建议和拆分令牌搜索被结合使用以将搜索结果返回给用户。 根据每个步骤的搜索查询结果的结果的数量和相关性,将结果合并或舍弃到最终结果集中。 分裂符号搜索适于从搜索查询的令牌生成两个拆分令牌,预期搜索令牌是拼写错误的。 由于拼写错误的位置未知,拆分令牌搜索扩大了响应搜索而提供的结果的范围。 在一个实施例中,分裂令牌搜索包括执行匹配前缀拆分令牌的令牌的前缀搜索和与后缀拆分令牌匹配的令牌的后缀搜索。 在一个实施例中,索引被特别地适于允许更有效地执行后缀搜索。