A User-based Personalized Recommendation Model For Gas Appliance

With the implementation of the national coal to gas strategy,gas products are facing a new round of development opportunities,with broad market demand.In the era of unprecedented development of e-commerce,shopping on Internet has been involved in a variety of products,including gas products.In this context,gas online sellers are also facing fierce competition in the market.For e-commerce marketing,to provide users with personalized recommendation is to improve their own core competitiveness and a favorable method to attract customers.But at present,online sales of gas manufacturers to provide product recommendations are based on the price of the product or the filter function,does not base on user’s personalized recommendations.Therefore,there is a demand for personalized recommendation function to online sellers of gas appliances.At present,there are many methods to solve the personalized recommendation problem in e-commerce,among which collaborative filtering recommendation method,content-based recommendation method and association rule-based recommendation method are the most popular.In addition,in order to solve the personalized recommendation problem more effectively,scholars also make continuous improvement and innovation on the basis of these methods,such as integrating the ideas of clustering,trust mechanism and big data analysis.However,there are still some problems in the research on personalized recommendation,among which data sparsity,cold start and method extensibility are the prominent problems.The purpose of this study is to help gas manufacturers to conduct more targeted marketing,and put forward a personalized recommendation model for gas manufacturers online.This model makes use of the user-based collaborative filtering idea to make recommendations,and adds user attribute information when calculating similar users,so that the model is also applicable to new users with only registered information.At the same time,this research model divides the gas appliances into “with sales records” and “no sales records”,which solves the cold start problem of gas appliance products.Finally,this paper conducts a case analysis of the model on the simulated data set,and verifies the feasibility of the model by comparing it with the traditional recommendation method,and then proves the effectiveness of the model through the experiment of sample set and test set.