Research On Key Techniques Of Cross-domain Recommendation Based On Personality Information

With the rapid increase of Internet information,in the Internet era where information is exponentially increasing,the cost and cost of obtaining useful and valuable information through the network is getting higher and higher,and the acquisition of information is becoming more and more difficult,which has also spawned the current From the current point of view,most websites and platforms use intelligent recommendation technology to provide information services to customers,so that customers can quickly find the information content they want,but from actual use.In view of the situation,the recommendation accuracy and user satisfaction of the current information recommendation system still need to be further improved.This is also the long-term demand for the growth of Internet information in the future.In this context,this paper considers from the perspective of personality information.Incorporating personality information into the design of user information recommendation system,based on the construction and design ideas of cross-domain recommendation system,based on the user information of social network platform,the algorithm design of cross-domain recommendation on e-commerce platform is realized.Effectively improve the accuracy and efficiency of crossdomain recommendations,the main research content of This thesis Next:1)The basic structure of the cross-domain recommendation system,the deep learning technology and the Big Five personality theory are studied,and the basic structure of the crossdomain recommendation system is analyzed.The recommendations in the cross-domain communication are inaccurate and cannot be applied to new users and the cold start problem of the new project proposes the introduction of Big Five personality information and deep learning technology solutions;2)Completing the design of user personality information perception and matching algorithm for social platform,aiming at the potential impact of personality information and determining the characteristics of user behavior and decision-making,by analyzing the personality information of users on social platforms,to further understand The algorithm of user behavior has completed the design of user personality determination method based on Big Five Personality.The theoretical model of BFM score table,personality factor value calculation and user personality score is given.Finally,the automatic matching of social platform user personality is completed.The implementation of the algorithm,through which the personality perception and measurement score of the social platform users can be realized.3)The design of the deep learning cross-domain recommendation algorithm based on user personality information is completed.The algorithm is improved based on the traditional collaborative filtering recommendation algorithm,introducing the Big Five personality factor,and then it is cold enough to adapt to new users and new projects.Start the problem,introduce deep learning,learn and train to better adapt to the processing of new users and new projects,complete the design of collaborative filtering module that introduces personality information,design of cross-domain deep learning model,and apply DNN network on this basis.The entire recommendation algorithm is implemented,which effectively improves the accuracy and adaptability of the recommendation.4)The experimental and simulation platform was built,and the experimental research was carried out.The performance of different algorithms was compared and analyzed,and the correctness and feasibility of the algorithm designed and constructed were verified.