The Research And Implementation Of Information Recommendation System Based On User Personality Data

As mobile Internet has developed rapidly,mobile intelligent terminals play an important role in our daily life.Mobile Apps based on location information have built a connection between the physical and the digital world to dig the potential value of the user personality data,such as location information and user behavior on the history.Mobile Apps based on user personality data have become a hot areas recently.So,this paper proposes a suite of information recommendation algorithm and implementation an information recommendation system based on user personality data.1.The most common information recommendation algorithm is a collaborative filtering recommendation algorithm based on neighborhood.But CF algorithm only considers part coefficient and does not take into account other important factors such as the location,time,or history data.Based on this situation,a new recommended algorithm named RABS was proposed,which take the history data of user,location,time and so on.Compared with traditional recommendation algorithm,RABS algorithm can more precisely dig out the potential demand,not only improve the quality and effect of recommendation service,but also help to eliminate information overload.2.A suit of recommendation system was developed,the prototype system includes:(1)the Android App subsystem,the App can collection the context information,such as time,location,in real time.It will update recommended list according to the history data of user consumption habits,the location and the time,and so on.(2)a business management subsystem in Web version:this subsystem is mainly to complete the inventory management,order management and commodity management,etc.(3)the background maintenance subsystem:include system maintenance,parameter settings,user management,category management,user rights management and report management,etc.(4)recommendation subsystem based on the Mahout:this subsystem using machine learning technique to separate the calculation of similarity and the user real-time scene information,so to save computation time and improve the system feedback speed.At present,the RABS recommendation algorithm has been used in the prototype system.After a series tests,the prototype system was proofed that can recommend commodities or services according to the personality data such as location and time,and the test results agree with the expected results.