Research And Application Of Personalized Recommendation Algorithm For Real Estate Information

Personalized information recommendation service,as a new information service mode emerging under the network environment,represents the development direction of information service in the future.At present,collaborative filtering technology has been successfully applied to the personalized recommendation system in many fields,such as e-commerce,social network,online music and real estate information services.This paper focuses on the field of real estate information,on the basis of in-depth study of various technologies in the field of personalized retrieval and recommendation,around the traditional synergy.The filter technology is faced with the real time,extensibility,accuracy and data sparsity,combining the features of the project content and user behavior,combining the clustering and combination algorithm ideas to improve the collaborative filtering algorithm,and proposes a collaborative filtering method as the framework and the fusion of content filtering groups.A recommendation algorithm is proposed and a housing intermediary intelligent management system is designed to verify the effectiveness of the algorithm.1.The literature found that in the data preprocessing stage,the traditional collaborative filtering algorithm overly depended on the user project score matrix,resulting in serious data sparsity and cold start problems.This paper makes use of the features of the project content and the user's behavior effectively.Through the method of feature combination and feature expansion,this paper excavate the long-term preference of the user,and then improve the recommendation quality of the recommendation system.2.In order to solve the problem of poor real-time and poor scalability in traditional collaborative filtering,this paper introduces a clustering model and proposes a new idea based on two-way clustering of user and real estate information,which combines the property of property information and user behavior,and sets the user and property information together to reduce the search for adjacent users or projects.Finding space not only improves the efficiency and real-time performance of the algorithm,but also supports the expansion of the system to a certain extent.3.Aiming at the problem of poor data sparsity and low accuracy in traditional collaborative filtering,this paper presents a combination recommendation algorithm based on content filtering and collaborative filtering,which combines the content of the project and the user behavior.Firstly,based on the content filtering of the project clustering,the evaluation of the non scoring items is evaluated,and the original user project score matrix is filled to solve the data sparsity and cold start problem.Secondly,on the basis of the completed user scoring matrix,the results of the project clustering,the property information content attribute and the user's line are fused.In order to improve the accuracy of the algorithm,a new fusion similarity computation method is proposed.Finally,collaborative filtering recommendation based on user clustering is adopted.Finally,based on the experimental data set of real real estate intermediary platform,the experimental simulation and test of the improved algorithm proposed in this paper are carried out,and the experimental results are compared with the traditional collaborative filtering recommendation algorithm.The experiment shows that the improved algorithm proposed in this paper improves the real-time performance and expansion of the recommendation system to a certain extent.The user's interest preferences are deeply mined through the user behavior characteristics and the content attributes of the real estate information,which effectively improves the accuracy of the recommendation algorithm.In this paper,a prototype system of housing agency is developed and applied to the system.The feasibility and effectiveness of the improved algorithm are verified.