Research On Sequence Recommendation Algorithm Based On Interpretabilit

With the progress of information technology,the internet has been closely linked to our daily lives.People engage in various activities through the internet every day,such as watching movies,shopping,browsing short videos,watching news,and listening to music.However,as the amount of information on the internet increases,it is difficult for people to find the most suitable information for themselves.The recommendation system analyzes users’ historical behavior to understand their preferences,and actively recommends information that interests them,meeting their personalized recommendation needs.Sequential recommendation is a new issue that has received much attention in recent years.By analyzing and modeling information such as user continuous behavior,user interaction with items,user preferences,and item popularity,it can more accurately depict user intentions,goals,and consumption trends of items,achieving more accurate,personalized,and dynamic recommendations.Explainable recommendation is one of the hot topics in internet applications such as social media,e-commerce,and content sharing.Explainable recommendation refers to solving the problem of recommendation reasons through personalized recommendation algorithms,that is,it not only provides recommendation results to users,but also explains why these items are recommended to users.Its main logic is to present the reasons for the recommendation results to users in a certain form,such as user or product similarity,user preferences for a certain product or feature,etc.These can all enhance user acceptance of the recommendation results.By explaining how and why a product is recommended to users,the transparency and interpretability of the recommendation system can be enhanced,which can help users make better and faster decisions.On this basis,this article intends to study a sequence recommendation algorithm based on interpretability,aiming to provide users with more accurate and interpretive recommendations for the next project.This paper mainly studies sequence recommendation algorithms based on interpretability,and proposes three methods to improve sequence recommendation,namely:(1)This article takes into account multiple attribute features of users,project sequences,and projects when embedding,and models multiple factors to increase the interpretability of the model.At the same time,this article also uses the idea of filtering algorithms to adaptively and effectively reduce data noise by using learnable filters.(2)This paper reorders the user’s historical item list and initial sorted list,not only capturing the interaction information between item item and item user,but also capturing the interaction information between user initial sorted list and user historical item list,so as to provide users with more accurate recommendations based on the above points.(3)This article takes into account the time factor in the model,modeling the longterm and short-term interests of users,and simultaneously considering the above two points in the model to improve the accuracy of the recommendation model.