Research On The Application Of Information Recommendation Based On UFDDR Algorithm

The problem we faced to be solved is now the Internet interconnection network a lot of information explosion era,massive database data to us provides a lot of valuable information,and how we can use existing technology to use these.data.In some shopping website according to your purchase history of goods and shopping cart commodity data,regularly recommended to the list of goods to buy what we want,and give the comprehensive index price and favorable rate etc.the best goods,make your shopping easy and reasonable;the application of the collaborative filtering recommendation algorithm,although this technology in the shopping website earlier application and one of the most successful algorithms,but the recommended value of accuracy in dependence on historical data and the strong recommendation data deviation is large,especially the recommended value of the effect is not obvious in the new buyers,so if the recommendation algorithm is applied in the fusion time deviation in algorithm(UFDDR),Accuracy will be greatly improved,in this paper,through the comparison of three algorithms,we can see that the algorithm in the RMSE,Precision,Recall and other indicators in the scoring system has a good effect on the application.Although researchers at home and abroad have a lot of research on the recommendation algorithm,but the study of the algorithm in the actual industry and commerce is not perfect convergence.