Research On The Personalized Recommendation Model Of Fund Based On Multi Strategy

Since 2013,have been tepid in the fund industry in the rapid development of the Internet led to Fujian,etc.in the market has made remarkable record,the number is rapidly growing size of the Fund,and with the continuous increase of per capita income,people gradually aware of the main financial management,began buying funds,stocks and other financial products.People are faced with thousands of funds,a variety of news and information,real-time stock quotes to feel dizzy,and thus not start.Therefore,how to tap the user preferences,to provide users with personalized recommendations and services to find their own demand for financial products become various fund sales platform urgent problems.And the current domestic fund sales platform,most have not personalized recommendation feature.For the above situation,the paper proposes a fund personalized recommendation model based on multi-strategy.In the model,the face of the new user,the new fund,in general,three different scenarios,we use three different strategies recommended and recommended the formation of the results by weighting and combining policy switch.In addition,to solve the problem for new users,this paper analyzes the new user buying behavior,proposed recommendations are based on user clustering algorithm popular fund.Finally,in order to verify multi-strategy recommended model can be applied,this paper based on the model to achieve the recommended system,and the method of A-B test,statistical user's click-through rate and purchase rate,which proves the effectiveness of recommended models.Compared with the traditional popular recommendation,the model recommended in this paper based on user characteristics,personalized recommendations fund interest to the user.Compared with simple personalized recommendation method,the effect of the model recommended in this paper is better,has improved to some extent on the recommended results and the accuracy of the conversion rate.