Hybrid recommendation method based on user commodity portrait and potential factor feature extraction

Abstract

The invention provides a hybrid recommendation method based on user commodity portrait and potential factor feature extraction. The hybrid recommendation method comprises the following steps of S100,extracting explicit feature representation of a user and a commodity through information of the user portrait and the commodity; s200, mapping the user and the commodity to a potential space to obtainpotential factor feature representation of the user and the commodity; and S300, performing feature extraction on the explicit features and the potential factor features by using a stack type noise reduction auto-encoder to obtain low-dimensional feature representation with higher robustness. According to the invention, the explicit feature space and the potential factor feature space of the userand the commodity are considered at the same time; two feature spaces are comprehensively considered, the defect of a single recommendation model is overcome, the problem of cold start of articles issolved, meanwhile, SDAE is adopted for extracting high-dimensional features, the problem of dimensionality disasters is effectively avoided, and due to the fact that random noise is added in the training process, the robustness of the algorithm is greatly improved.

Description

technical field [0001] The present invention relates to feature representation and cleaning of text information in user and product portraits, and a feature extraction method using score matrix decomposition, especially a hybrid recommendation method based on user product portraits and latent factor feature extraction, belonging to field of artificial intelligence. Background technique [0002] The recommendation system is a model that can predict users' preferences or ratings for products and information, and is an important channel for users to find information and products suitable for them from the vast amount of information on the Internet. Use the user's historical behavior of the product (such as: purchase, browse, click, evaluation, etc.) to mine each user's preferences, and then make personalized recommendations for users. The historical behavior of users on products is usually divided into two categories: one is explicit feedback, that is, the rating information o...

AI Technical Summary

Problems solved by technology

[0012] Compared with the traditional recommendation method, such as the collaborative filtering method, this method encounters the situation that a new product is added to the scoring matrix, that is, when a new product appears, no user has evaluated it. Recommendations cannot be made because there is not enough scoring data. Using user and product portraits to establish user and product characteristics can effectively solve the cold start problem that collaborative filtering has for new users and new products. The method of user and item latent factors can map users and items into their respective latent spaces, and describe the feature representation of users and items through latent factors, which can describe the attributes of users and items from multiple latent spaces and overcome the problems based on The problem of sparse features in the method of user and product portrait, but also because of ignoring the inherent feature attributes of users and products, the recommendation efficiency will decrease

Images