专利名称：Gradient based optimization of a ranking
measure
发明人：Olivier Chapelle申请号：US11863453申请日：20070928公开号：US07895198B2公开日：20110222
专利附图：
摘要：Methods, systems, and apparatuses for generating relevance functions forranking documents obtained in searches are provided. One or more features to be usedas predictor variables in the construction of a relevance function are determined. The
relevance function is parameterized by one or more coefficients. A query error is definedthat measures a difference between a relevance ranking generated by the relevancefunction and a training set relevance ranking based on a query and a set of scoreddocuments associated with the query. The query error is a continuous function of thecoefficients and aims at approximating errors measures commonly used in InformationRetrieval. Values for the coefficients of the relevance function are determined thatsubstantially minimize an objective function that depends on the defined query error.
申请人：Olivier Chapelle
地址：Mountain View CA US
国籍：US
代理机构：Fiala & Weaver P.L.L.C.
更多信息请下载全文后查看
因篇幅问题不能全部显示，请点此查看更多更全内容