Research On E-commerce Of Agricultural Recommendation System Based On Users' Perspective

With the Central No.1 file put forward the "'Internet +' agriculture" action plan,e-commerce of agricultural in China quietly rising,breaking the traditional agricultural market "face to face" marketing model,to achieve the real estate transactions of agricultural products.There are many types agricultural products.The agricultural market is widely distributed,and with the seasons,the area changes.While being affected by natural conditions,Crops are produced in a lumpy distribution.In different regions or seasons,Users should plant different types of crops,Users' personalized requirements are obvious,But the current e-commerce of agricultural systems constantly expand the scope of agricultural products business,causing the user to fall into the information redundancy environment and on the other hand the system's personalized service performance is poor,recommended logic is single and lack of targeted recommending strategy.Users can not get applicable agricultural products timely and accurately.It is necessary to design an efficient and efficient e-commerce of agricultural systems.In this paper from users' perspective,being guided by users' appeal and experience,design three agricultural products recommendation model for registered users,non-registered users and new users respectively,achieve an e-commerce of agricultural recommender system based on user perspective.Details are as follows:For registered users,through analysis characteristics of user and agricultural product,we can concluded that user and product characteristics are important factors to determine the user's demand,so we propose two concepts which are user's portrait and number of user's rating on the same category of products,calculating user's portrait similarity and user's rating difference,combined with user rating matrix,we proposed a multi-dimensional user similarity,aiming at the limitation of similarity calculation method in traditional user-based collaborative filtering recommendation algorithm,achieve an e-commerce of agricultural recommendation algorithm Based on Registered User Portrait,this method id applied to the actual business platform,good application results have been obtained.For non-registered users,through tap the user history traces,we extract the user intent label,adopt user' intention characteristics and agricultural product characteristics we built user's intention model and product characteristics model,based on the idea of targeted advertising,guess user's potential intent,matching user's intention model and agricultural product model,Finally,push high match degree products to users in the form of advertising.For new users,we employ some substituting strategy,such as hot products,high evaluation or discount products.So from users' perspective,combined recommended technology with characteristics of users and agricultural products,design and implementation an efficient e-commerce of agricultural systems,can meet users' needs,stimulate consumer spending and promote the flow of agricultural products.