Research On Collaborative Recommendation Algorithm Based On Explicit And Implicit Feedback

In most online shopping platforms,users' historical behavior data(e.g.,ratings,likes and browsing history)implicitly contain user preference information.Many recommender systems regard it as an important source of input data.According to the feedback mechanism of users,users' online behavior data can be divided into explicit feedback data and implicit feedback data,and for the current E-commerce field,they both coexist,and seldom exist explicit feedback or implicit feedback data.However,most of the previous recommendation algorithms focus on explicit or implicit feedback alone.It is always a difficult point to study how to integrate explicit implicit feedback for the recommended tasks.Merely using explicit feedback data(e.g.,ratings,ratings)often ignores user-like similarities,and recommendations based only on implicit feedback data(e.g.,browsing history,clicks,shopping carts)ignore user preferences for the product.So some traditional recommendation algorithms cannot meet the urgent demand of the current E-commerce field for the recommended technology.This paper mainly studies and solves the problem of collaborative ranking recommendation based on explicit implicit feedback.Starting from the perspective of how to combine explicit implicit feedback data,we propose a collaborative recommendation model EIPM based on implicit explicit feedback,while the implicit feedback for further research.How to combine the semantic information of structured data contained in the knowledge base with the implicit feedback of users is the key point to be explored in this paper.Finally,we propose a collaborative ranking recommendation algorithm KE-BPR based on structural features and implicit feedback.For the new recommended models EIPM and KE-BPR algorithm proposed in this paper,we show the recommended effect through sufficient experiments based on two real data sets.The main work and achievements of this paper are as follows:(1)In this paper,a collaborative ranking recommendation algorithm KE-BPR based on feature and implicit feedback is proposed,which makes full use of the semantic information contained in the knowledge base and uses the knowledge base embedding method to extract the feature representation of structured knowledge from the knowledge base to enrich the recommended items.Then,we integrate the implicit feedback between users and items and the structural embedding of knowledge base,and further improve the performance of recommendation algorithm through collaborative joint learning.(2)We propose an explicit implicit feedback model(EIPM)based on explicit implicit feedback(EIPM).Firstly,based on the implicit feedback data of users,we select the items most likely to be viewed by the user from the set of items to be sorted.Explicit feedback data is used to predict the user's rating of the item,and the sequence of prioritized items is rearranged to further enhance the effect of the recommended algorithm.(3)For the newly proposed EIPM and KE-BPR algorithms proposed in this paper,we conduct extensive experiments based on two real data sets to verify their respective validity.Experimental results show that KE-BPR,which makes full use of knowledge of the structure in the knowledge base,and EIPM,a collaborative recommendation model with explicit implicit feedback,are superior to other baseline algorithms in recommendation effect.