Deep learning-based cate store recommendation method

Abstract

The invention discloses a deep learning-based cate store recommendation method. The method comprises the following steps of: obtaining comment data of a first user; training the comment data through aconditional random field model and a word2vec algorithm so as to obtain an entity label, training the comment data through an emotion analysis algorithm so as to obtain emotion judging data and obtaining a label of a store according to the entity label and the emotion judging data; and obtaining input data of a second user, recognizing an intention of the second user according to the input data and a preset deep learning algorithm, and recommending the label to the second user when the intention of the second user is the recommended store so as to complete the recommendation of the store. According to the method, emotion analysis can be accurately carried out user comments to ensure the entity clustering correctness, and the intentions of the users can be correctly and automatically judged so as to ensure that recommendations in chat contexts as inputs of labels.

Description

technical field [0001] The invention relates to the field of food recommendation, in particular to a method for recommending food stores based on deep learning. Background technique [0002] In various recommendation fields, the more mature algorithms on the market include: user- or item-based collaborative filtering algorithms, user-item matrix decomposition / recommendation algorithms based on graph models or relational networks. These algorithms are powerful and useful when there is a large amount of data about user interactions with items. However, in many cases, we cannot obtain the interaction data between a large number of users and the items to be recommended, and it is likely that there are only unilateral data of users or items. If there is only one-sided data of users, in general, only one-sided feature attributes of users or items can be mined, which only serves as a model for rough classification of targets and matching with each other. However, the effect of th...

AI Technical Summary

Problems solved by technology

[0004] The technical problem to be solved by the present invention is how to realize the recommendation in the field of food, the accurate judgment of automatic intention in the context, and the recommendation of precise food stores when a large amount of user item interaction data cannot be obtained

Images