Study On The Personalized Recommendation Method For Product Combination Based On Collaborative Filtering

When online shoppers purchase multiple products in a single transaction,the information about the products and shops they are facing is more than when buy only one product,this may cause the online shopping users directly to abandon the deal because of the heavy work of information extraction,so online shopping like that without a correct personalized recommendation measure usually bring a larrnumber number of customers losing,and it will hold back e-commerce further.The traditional product personalized recommendation method pays less attention to this case of purchasing multiple products in a single transaction,in the product personalized recommendation list which is returned to online shopping users,most of them are just for one single product without the best scheme for purchasing,thus,the motivation of online shopping users will be reduced when they buy multiple products in a transaction and reduce the success rate of recommendation of personalized recommendation system further.For that reason,it is necessary to study on the personalized recommendation problem considering the promotion strategy for product combination and explore a new product combination personalized recommendation method.This method can recommend a series of product combination to users based on their interests and offer them the best purchasing schemes for each product combination,to the greatest extent to enhance the online shopping enthusiasm of users and improve the recommendation effect of personalized recommendation system.This paper deeply studied the personalized recommendation method considering promotion strategy for product combination based on the summary and analysis of the traditional product personalized recommendation methods.Now this paper is specifically described from the following four aspects of the work:(1)Proposed a kind of product combination recognition method based on the similarity of items.Aim at the problem of information overload that the online shopping users are facing when buy multiple products in a deal,this paper proposed a kind of product combination recognition method based on the similarity of the items.Firstly,through the item similarity calculation method,mumbert the item similarity matrix,then find out the preliminary products which are the most similar to the product that is clicked by the user now,then find out the products which are the most similar to the products which are find out in the previous step,lastly,the products are combined,and they are the product combinations which are recognized.(2)Proposed a kind of prediction method of rating for product combination based on matrix factorization.Aim at the product combination TOPN sorting problem based on users'interest,this paper proposed a kind of prediction method of rating for product combination based on matrix factorization.Firstly,through the recommendation method which is based on matrix factorization,forecast the rating for each product which is belong to the product combinations,then give the weight to each product in the combination,then calculate the combination' rating weighted sum,and it is product combination' comprehensive rating.(3)Proposed a kind of purchasing scheme rnumberneration method considering promotion strategy for product combination.Consider that there are many schemes for each product combination,users need to make a choice for buying the product combination,and this is a hard work for them,aim at the problem,this paper proposed a kind of purchasing scheme rnumberneration method considering promotion strategy for product combination.Firstly,build the model based on the minimum payment cost,for improving the model further,consider the promotion strategies:freight,discounts,coupons,gifts.Lastly,proposed the method for solving the model and achieve the best purchasing scheme for product combination.(4)Showed the application of the personalized recommendation method.Take the personalized recommendation problem for book combination based on the Taobao e-commerce platform as an example,this paper showed the application of the personalized recommendation method considering promotion strategy for product combination and proved that it is reasonable and practical.This method can help the users achieve the product combination they interested in and the purchasing scheme for it quickly.It laid a foundation for the related research.