Knowledge graph entity feature mining method based on translation model

• Main Author: XIAO QINGLIN
• Format: Patent
• Language: Chinese; English
• Published: 05.11.2019
• Subjects: CALCULATING; COMPUTING; COUNTING; DATA PROCESSING SYSTEMS OR METHODS, SPECIALLY ADAPTED FORADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORYOR FORECASTING PURPOSES; ELECTRIC DIGITAL DATA PROCESSING; PHYSICS; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE,COMMERCIAL, FINANCIAL, MANAGERIAL, SUPERVISORY OR FORECASTINGPURPOSES, NOT OTHERWISE PROVIDED FOR

A knowledge graph entity feature mining method based on a translation model is characterized by comprising the following steps of redefining an energy function of the translation model through a new structure information vector to obtain a new energy function, and obtaining the new translation model through training of the new energy function; based on the new translation model, obtaining a firstgroup of knowledge information and a plurality of entities, screening a plurality of associated high-frequency vocabularies from the first group of knowledge information of the labeled entities, and screening a plurality of word groups with the highest co-occurrence frequency in the plurality of high-frequency vocabularies to generate an ordered association rule; and obtaining a second group of knowledge information of the current domain, mining the second group of knowledge information according to the ordered association rule, and filtering the plurality of entity candidates according to a predetermined filtering ru