Research On Recommended Algorithm And Implementation Of Exchange Goods Network

With the rapid development of the Internet,people can get large amounts of data.As the increasing amount of data,how to capture the information of interest in massive data has become one of the hot topics of large data analysis.In this case,the personalized recommendation technology came into being under this circumstance.Personalized recommendations can improve the user efficiency of finding the information that interested by the user during the valid time.on the other hand,they can make the business initiatively provide the useful information to users in time.Therefore,the study of personalized recommendation algorithm has great commercial value and significance.It has aroused widespread concern and research in academia and the business community.Hence,this thesis studies focus on the user-based personalized collaborative filtering algorithm and apply the algorithm to the construction of the exchange network platform.The main research work includes:(1)For the data sparsening of cooperative filtering algorithm,a filling algorithm based on project activity is proposed.The algorithm uses the way of slope one to preload the user's scoring data,which effectively solves the problem of using the number of single user scores to calculate the sparseness of user similarity data.The filling method is simple and reasonable.Compared with the traditional filling method,the proposed algorithm can enhance the sparseness of data and improve the accuracy of user similarity calculation.(2)A collaborative filtering algorithm based on distance penalty factor is proposed to improve the sparse similarity calculation accuracy of collaborative filtering algorithm.This algorithm corrects the traditional Pearson similarity by using all the scoring distances of the common score among the users as the penalty factor.By adjusting the similarity to the distance penalty factor,the user similarity is adjusted adaptively and the user similarity accuracy is improved in the collaborative filtering optimization algorithm.The experiment proves the validity of the proposed algorithm.(3)In view of the practical problems of children's picture-making transactions,the exchange goods network platform is constructed and the above-mentioned collaborative filtering algorithm is integrated into the practical platform.The platform mainly includes four modules:recommended books module,book transaction module,member management module,book maintenance module.The algorithm proposed in this thesis has been successfully applied to the exchange goods network platform,and has realized the basic function of the recommendation module.