Research And Application Of Spatial-temporal E-commerce Drug Recommendation Algorithm

With the proposal of the slogan "Internet + medicine",both the types and quantity of drugs on e-commerce platforms have witnessed rapid growth.Novel Coronavirus outbreak in 2020 makes people see the value of pharmaceutical e-commerce.Pharmaceutical e-commerce has become a future trend and has a great development space in China.However,faced with the huge amount of drug information on the Internet,consumers often do not know how to make a choice,and it is difficult for them to find the right drug quickly.On the other hand,the pharmaceutical e-commerce websites generate a huge amount of sales records every day,and these data are properly utilized.In order to solve these problems,it is urgent to introduce personalized recommendation system and use consumers' previous consumption data to solve the problem of drug overload.The traditional recommendation algorithm does not consider the influence of time and space factors on the recommendation system,and the existing recommendation algorithm based on time and space does not fully consider the characteristics of drugs,which often results in unsatisfactory recommendation results.This article will give full consideration to the influence of time and space for drugs,the change of user interest and drug combination of seasonal,regional characteristics,into the traditional collaborative filtering recommendation algorithm,put forward the collaborative filtering recommendation algorithm based on space-time,satisfy the user's preferences,can also help the users to make decisions,to improve the quality of the recommendation of drug.The main research contents are as follows:(1)Considering the influence on the recommendation system from the time dimension,a time based collaborative filtering recommendation algorithm was proposed.Firstly,according to the characteristics that the user's interest will change with time,the time attenuation function is introduced.The seasonal weighting function is introduced for the seasonal variation of drugs.Finally,the time factor is incorporated into the collaborative filtering recommendation algorithm to improve therecommendation quality.(2)Proposed a spatial-temporal collaborative filtering recommendation algorithm.On the basis of time-based collaborative filtering recommendation algorithm,the spatial information of users is integrated into the recommendation algorithm,so that the influence of time and place changes on user preferences is considered in the process of recommendation.The algorithm based on collaborative filtering recommendation algorithm based on time,in view of the spatial information users,the pyramid model,will all users into the each layer of the pyramid node,then in each layer using the collaborative filtering recommendation algorithm based on time to get local recommendation,the final will be local recommendation results obtained shall be carried out in accordance with the weighted linear sum of the final recommendations as a result,in order to improve the prediction accuracy.(3)The collaborative filtering recommendation algorithm based on time and space proposed in this paper is applied to the medical e-commerce website,and a personalized recommendation system for drugs is designed and implemented to realize the function of personalized recommendation for drugs.This system can solve the problem of cold start of new users,improve the satisfaction of users to the system platform,retain the old customers of the system,and promote the sales volume of drugs,which is of great significance to promote the development of medical e-commerce.