Enterprise service recommendation method based on enterprise feature propagation

Abstract

An enterprise service recommendation method based on enterprise feature propagation performs feature propagation on enterprises in an interaction record by using a knowledge graph associated between enterprise features and a target service, automatically mines enterprise association paths, depicts associated features between the enterprises, and combines the associated features with the enterprise features. And the new loss function is combined with deep learning to obtain the interaction probability of the enterprise and the service, so that the problem of poor recommendation effect caused by deviation from the relationship between the enterprises due to the fact that a general framework only uses interaction data and basic information can be solved, and accurate recommendation of the service scheme of the enterprise is realized. Through automatic mining of relationships between enterprises, discovery of association paths between the enterprises, interaction prediction scoring of enterprise features, and feature propagation of the enterprises in interaction records by target services and an interaction framework, the problem that the enterprises are difficult to select service schemes is solved.

Description

technical field [0001] The invention relates to the technical field of deep learning, in particular to an enterprise service recommendation method based on enterprise feature propagation. Background technique [0002] With the rapid development of the economy, small, medium and micro enterprises have become an important force in the current national development. They play an irreplaceable role in improving people's livelihood, promoting employment, and stimulating the economy. They have rapidly changing businesses and imperfect enterprise service supply systems. Enterprise services play an important role in the development of small, medium and micro enterprises. In the face of massive enterprise services, it is difficult to accurately find suitable enterprise service solutions, so the intelligent recommendation of enterprise service solutions is particularly important. [0003] In previous recommendation systems, the main traditional recommendation algorithms have been wide...

AI Technical Summary

Problems solved by technology

However, in these recommendation algorithms, ordinary interaction information and auxiliary information are used, but in enterprise service recommendation, it is difficult to map the characteristics of enterprises and the associations between enterprises.

The workload of manually designing the meta-path associated with the enterprise is extremely heavy, and it is difficult to dig deeper into the association relationship. The complex business needs of the enterprise cannot be fully reflected, making it impossible to mine the potential services required by the enterprise.

Images