CN-115237966-B - Food path recommending method and device, electronic equipment and storage medium

CN115237966BCN 115237966 BCN115237966 BCN 115237966BCN-115237966-B

Abstract

The application relates to the technical field of computers and artificial intelligence algorithms, and provides a food path recommending method, a food path recommending device, electronic equipment and a storage medium. The food path recommending method comprises the steps of obtaining a candidate path set for recommending target food to a target user based on a food recommending request of the target user, determining a target recommending path based on the candidate path set, and recommending food based on the target recommending path, wherein different candidate paths in the candidate path set are used for representing different recommending reasons for recommending the target food. The method provided by the application can be used for realizing the purpose of flexibly and pertinently pushing the recommended path while recommending food to different users, and improving the accuracy and the interpretability of the recommended path of the food.

Inventors

LI YUXIAO
Song Yunjin
Cao tianyuan
MOU XIAOFENG
TANG JIAN

Assignees

美的集团(上海)有限公司
美的集团股份有限公司

Dates

Publication Date: 20260508
Application Date: 20220713

Claims (12)

1. A food path recommendation method, comprising: Based on a food recommendation request of a target user, acquiring a candidate path set for recommending target food to the target user, wherein different candidate paths in the candidate path set are used for representing different recommendation reasons for recommending the target food; Determining a target recommended path based on the candidate path set; food recommendation is performed based on the target recommendation path; The acquiring a candidate path set for recommending target food to the target user based on the food recommendation request of the target user comprises the following steps: determining a food recommendation result for a target user based on a user-food preference model and a food recommendation request of the target user, wherein the user-food preference model is a model obtained by training historical operation record data of different users on different foods; acquiring a food knowledge graph, wherein the food knowledge graph represents graphs of relationships among related users, food and cooking equipment from different dimensions; acquiring an entity relation confidence coefficient map corresponding to the food knowledge map, wherein the entity relation confidence coefficient map is a map determined after the food knowledge map is subjected to entity relation quantification treatment; and determining a candidate path set for recommending the target food to the target user based on a food path reasoning model and the food recommendation result, wherein the food path reasoning model is a model obtained by training an initial recommendation model based on the entity relation confidence level map.
2. The food path recommendation method according to claim 1, wherein after the obtaining of the entity relationship confidence map corresponding to the food knowledge map, the method further comprises: and determining a candidate path set for recommending the target food to the target user based on the food path reasoning model, the food recommendation result and the food knowledge graph.
3. The food path recommendation method of claim 1, wherein said determining a target recommended path based on said candidate path set comprises: recall the candidate path set based on a recall policy, the recall policy characterizing recommending foods based on an association between users, an association between foods, an association between users and foods, and an association between food tags and foods; And sequencing and denoising the recalled candidate path set to determine a target recommended path.
4. The food path recommendation method of claim 3 wherein said sorting and denoising said recalled candidate path set to determine a target recommended path comprises: sorting and denoising the recalled candidate path set, and determining a path to be recommended; obtaining an audit scoring result for the path to be recommended; And correcting the path to be recommended based on the auditing scoring result, and determining a target recommended path.
5. The food path recommendation method according to claim 4, further comprising, after said obtaining an audit score for said path to be recommended: updating the user-food preference model based on the audit score.
6. The food path recommendation method according to claim 3, wherein after said food recommendation based on said target recommended path, said method further comprises: acquiring feedback information of the target user aiming at the target recommended path; and updating the recalled candidate path set based on the feedback information.
7. A food path recommendation method according to any one of claims 1 to 3, wherein said making a food recommendation based on said target recommended path comprises: And determining a visual recommended path description text for the target food based on a natural language description template and the target recommended path, and pushing based on the visual recommended path description text.
8. A food path recommendation method according to any one of claims 1 to 3, wherein said making a food recommendation based on said target recommended path comprises: Performing visual text description processing on the target recommended path based on a natural language generation model, determining visual recommended path description text for the target food, and pushing based on the visual recommended path description text, wherein the natural language generation model is a model obtained by training a neural network based on entity language tags of different corpus.
9. A food path recommendation method according to any one of claims 1 to 3, wherein the process of obtaining a food knowledge graph comprises: acquiring a user sub-map based on personal information and demand information of different users; acquiring a food sub-map based on food information and cooking information of different foods; acquiring equipment sub-maps based on the types and the categories of different cooking equipment; And carrying out spectrum coupling on the user sub-spectrum, the food sub-spectrum and the equipment sub-spectrum to obtain a food knowledge spectrum.
10. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements the food path recommendation method of any one of claims 1 to 9 when the program is executed by the processor.
11. A non-transitory computer readable storage medium having stored thereon a computer program, which when executed by a processor implements the food path recommendation method according to any one of claims 1 to 9.
12. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the food path recommendation method according to any one of claims 1 to 9.

Description

Food path recommending method and device, electronic equipment and storage medium Technical Field The application relates to the technical field of computers and artificial intelligence algorithms, in particular to a food path recommending method, a food path recommending device, electronic equipment and a storage medium. Background Along with the continuous improvement of living standard, the requirements of people on living quality are higher and higher, such as diet quality, so people often use a menu recommendation system to recommend a menu for themselves, but the menu recommendation system only can ensure the recommendation speed and the accuracy of recommended contents of the menu for menu recommendation, but cannot show the recommendation reason for recommending the menu to a user. Therefore, how to interpret recommended recipes to users becomes a research hotspot. Disclosure of Invention The present application is directed to solving at least one of the technical problems existing in the related art. Therefore, the application provides the food path recommending method, which realizes the purpose of recommending food to different users and simultaneously and flexibly pushing the recommended path in a targeted way, and improves the accuracy and the interpretability of the food recommended path. The application further provides electronic equipment. The application also proposes a non-transitory computer readable storage medium. The application also proposes a computer program product. According to an embodiment of the first aspect of the present application, a food path recommendation method includes: Based on a food recommendation request of a target user, acquiring a candidate path set for recommending target food to the target user, wherein different candidate paths in the candidate path set are used for representing different recommendation reasons for recommending the target food; Determining a target recommended path based on the candidate path set; and recommending food based on the target recommendation path. According to the food recommendation method provided by the application, the terminal equipment firstly acquires the candidate path set for recommending the target food to the target user aiming at the food recommendation request of the target user, and then determines the target recommendation path based on the candidate path set, and because different candidate paths in the candidate path set are used for representing different recommendation routes for recommending the target food, when the target food is a plurality of and at least two target foods are repeated, the target recommendation path can be determined from the candidate path set by carrying out processing modes such as de-duplication, filtering and merging on the candidate path set, thereby realizing the purpose of flexibly and pertinently pushing the recommendation path while recommending the food to different users, and improving the accuracy and the interpretability of the food recommendation path. According to one embodiment of the present application, the acquiring, based on a food recommendation request of a target user, a candidate path set for recommending a target food to the target user includes: determining a food recommendation result for a target user based on a user-food preference model and a food recommendation request of the target user, wherein the user-food preference model is a model obtained by training historical operation record data of different users on different foods; acquiring a food knowledge graph, wherein the food knowledge graph represents graphs of relationships among related users, food and cooking equipment from different dimensions; And acquiring a candidate path set for recommending target food to the target user based on the food knowledge graph and the food recommendation result. According to one embodiment of the present application, the acquiring a candidate path set for recommending a target food to the target user based on the food knowledge graph and the food recommendation result includes: acquiring an entity relation confidence coefficient map corresponding to the food knowledge map, wherein the entity relation confidence coefficient map is a map determined after the food knowledge map is subjected to entity relation quantification treatment; and determining a candidate path set for recommending the target food to the target user based on a food path reasoning model and the food recommendation result, wherein the food path reasoning model is a model obtained by training an initial recommendation model based on the entity relation confidence level map. According to an embodiment of the present application, after the obtaining the entity relationship confidence map corresponding to the food knowledge map, the method further includes: and determining a candidate path set for recommending the target food to the target user based on the food path reasoning model, the food recommendat