Method of expert combination recommendation based on knowledge map

Abstract

The invention discloses an expert combination recommendation method based on a knowledge map. The invention is mainly based on a constructed domain knowledge map, and combines a similarity idea and anassociation mining. The method comprises: to process the history expert review data, first using the FP-Growth to obtain probabilistic logic data to make knowledge map, and then eliminating the differences through similarity algorithm, extracting the similar items with the recommended content, and then extracting experts and combining them according to the rules, and designing an excellent methodto evaluate the combinations by frequent items and confidence according to the relational knowledge. The method of the invention effectively recommends a combination of graph auditing experts with the highest score, which improves the cooperative examination efficiency of the experts and increases the use value of the data of the historical item examination expert set.

Description

technical field [0001] The invention belongs to the technical field of recommendation algorithms and knowledge graphs, in particular to an expert combination recommendation method based on knowledge graphs. Background technique [0002] The expert combination recommendation algorithm in the present invention has an important function and significance for project-oriented expert group recommendation. When faced with the problem of recommending a combination in the recommendation system, researchers will choose to integrate it into the group recommendation system based on some group clustering, similarity, classification analysis, etc., or through user scoring matrix, semantic analysis and other technologies. How to explore the correlation between groups and projects, and the potential relationship between group individuals, so as to provide valuable relational knowledge for group recommendation, and optimize the group recommendation algorithm to provide efficient knowledge se...

AI Technical Summary

Problems solved by technology

[0016] Some traditional project-oriented expert group recommendation algorithms are manually selected, and some are only recommended according to the project and recommended objects. The artificial recommendation method consumes human resources and has a certain one-sidedness in the recommendation, so it cannot analyze the project and recommendation. Recommend the relationship between objects and make recommendations, resulting in insignificant recommendation effects, so they are replaced by recommendation algorithms based on the relationship between projects and experts. These methods cannot make full use of the relationship between experts, and cannot reach expert group recommendations accuracy, so that the recommended expert combination cannot meet the corresponding recommendation requirements

Images