发明名称 Semantic context based keyword search techniques
摘要 Provided are semantic context based keyword search techniques. A query with one or more keywords, one or more relevant words, and one or more selected services is received. A general relevance indicator for services in a category is calculated based on keyword relevance indicators of the one or more keywords in the query for each service among services in the category and relevant word relevance indicators of the one or more relevant words in the query for the services in the category. The services in the category are ranked based on the general relevance indicator for each service. In response to determining that at least one of the selected services is not ranked high with reference to the ranked services in the category, at least one of: one or more additional keywords, one or more additional relevant words, and one or more additional services to provide as recommendations are identified.
申请公布号 US9589050(B2) 申请公布日期 2017.03.07
申请号 US201414247054 申请日期 2014.04.07
申请人 International Business Machines Corporation 发明人 Tao Jim J.
分类号 G06F17/30 主分类号 G06F17/30
代理机构 Konrad, Raynes, Davda and Victor LLP 代理人 Davda Janaki K.;Konrad, Raynes, Davda and Victor LLP
主权项 1. A method, comprising: receiving, with a processor of a computer, a query with keywords that each identify a key concept, relevant words that capture a context of the keywords, and one or more selected services; calculating a general relevance indicator for each service among services in a category using:GRIQ=∑i=1Mqk⁢⁢KRIi+∑i=1Mqr⁢⁢αi2×RRIi((Mqk+∑i=1Mqr⁢⁢αi)⁢(∑i=1Msk⁢⁢KRIi2+(∑i=1Msr⁢⁢αi2×RRIi2)))12 wherein KRIi is a keyword relevance indicator of a keyword i, from the keywords, associated with a service of the services in the category, wherein Mk is a total number of the keywords in the query, wherein RRIj is a relevant word relevance indicator of a relevant word i, from the relevant words, associated with the service, wherein Mr is a total number of the relevant words in the query, and wherein αi is a weight for the relevant word i; ranking the services in the category based on the general relevance indicator for each service; and in response to determining that at least one of the selected services is not ranked high with reference to the ranked services in the category, recommending synonyms to be used to substitute corresponding keywords and synonyms to be used to substitute relevant words in the query;recommending new relevant words by: retrieving names of properties and names of relationships of a concept identified by a keyword among the keywords in an ontology database as candidate relevant words; andin response to determining that the candidate relevant words are not in the query and no synonyms are found, providing a recommendation that the candidate relevant words be included in the query as the new relevant words; andrecommending one or more additional services to be included in the query.
地址 Armonk NY US