A recommendation system and method based on the analysis of coupling relationship between users and items

Abstract

The invention discloses a recommendation system and method based on analysis of coupling relationship between users and items, including data collection and processing, division of data sets, construction of coupling models, training models and item recommendation. The present invention considers the very microcosmic coupling relationship between user features and item features. When the scoring information is relatively sparse, this coupling relationship can recommend items that users like, which improves the quality of recommendation; and uses the Attention mechanism to capture users' opinions on The preference degree of different features of the item makes the recommendation system have better recommendation effect and explainability. In addition, the present invention uses Doc2vec to extract explicit features of users and items from comment texts, which reduces the dimension of explicit features of users / items, speeds up model operation and improves recommendation accuracy. Compared with matrix decomposition, the present invention adopts The non-linearity of convolutional neural network and deep neural network helps to learn the interaction between features at a deeper level.

Description

technical field [0001] The invention belongs to the technical field of natural language processing and computer artificial intelligence, and in particular relates to a recommendation system and method based on the analysis of coupling relationship between users and items. Background technique [0002] With the rapid development and popularization of Internet information technology, people are more and more fond of shopping online and commenting on and scoring items online. However, facing so many similar items on the e-commerce platform, consumers have to spend a lot of time choosing on your favorite projects. Therefore, it is very important to recommend items that users like and have interpretability. Many existing recommendation systems have poor explainability and consider users and items to be independent and identically distributed, ignoring the heterogeneity and coupling between users and items. In fact, there are various coupling relationships between users and items...

AI Technical Summary

Problems solved by technology

Therefore, it is very important to recommend items that users like and have interpretability. Many existing recommendation systems have poor explainability and consider users and items to be independent and identically distributed, ignoring the heterogeneity and coupling between users and items. In fact, there are various coupling relationships between users and items, between user features, and between item features. This coupling relationship can better explain users' preferences for items.

Images