Method and device for constructing recipe recommendation strategy based on knowledge graph

Abstract

The invention discloses a method for constructing a recipe recommendation strategy based on a knowledge graph. The method comprises the following steps: obtaining health information and food taboo information of a user, and generating a user sub-graph G' corresponding to a user i according to a nutrition health knowledge graph G; performing representation learning on the nutrition health knowledgegraph G, and encoding the nutrition health knowledge graph G into a vector space; generating a user-recipe preference matrix R' according to the nutrition health knowledge graph G and the user sub-graph G'; performing matrix decomposition on the user-recipe preference matrix R' to generate the recipe candidate set Ci of the user i; and screening foods in the recipe candidate set Ci according to preset conditions to obtain a preferred recipe set, and recommending the preferred recipe set. The method disclosed by the present invention can realize the dynamic recipe recommendation strategy undermultiple conditions. The invention also discloses a device for constructing the recipe recommendation strategy based on the knowledge graph. The device can be used to construct the recipe recommendation strategy and solve the "what to eat" problem of the user.

Description

technical field [0001] The present invention relates to the field of knowledge graphs, and more specifically, to a method for constructing recipe recommendation strategies based on knowledge graphs. The invention also discloses a device for constructing a recipe recommendation strategy based on a knowledge map. Background technique [0002] With the continuous improvement of people's living standards. Whether eating at home, ordering takeaway or eating out, people can choose more and more delicacies every day. Faced with such a variety of delicacies, people don’t know how to choose, and the existing recipe recommendation strategies are mostly single-dimensional. Or it is a simple superposition of several dimensions according to the weight, without actually establishing the mutual influence relationship of different dimensions, so the existing recipe recommendation method cannot construct a dynamically changing recipe recommendation strategy under multiple constraints. [0...

AI Technical Summary

Problems solved by technology

However, although the current knowledge graph technology has developed relatively maturely in the field of search, it has not been widely used in other fields. There is no precedent on the market for using knowledge graphs to construct recipe recommendation strategies; Relational construction, high technical threshold, currently lacks a ready-made recipe recommendation model based on knowledge graph, so the corresponding recipe recommendation model needs to be built from scratch, and the recipe recommendation logic based on knowledge graph should be improved and sorted out

Images