| With the advent of the information age and the rapid rise and development of new generation technologies such as cloud computing and big data,agricultural informatization has become the driving force for the development of modern agriculture.Influenced by traditional agriculture,agricultural practitioners have little initiative to obtain agricultural information.Therefore,it is a very important task for agricultural practitioners to obtain effective agricultural information.Due to the large number of farmers and the huge amount of news resources,ordinary push technology will push all information to users,which will lead to users not interested in information,forming the opposite effect.Item-Based collaborative filtering algorithm can infer user preference information according to user’s browsing records,and can calculate the similarity between information.In view of this,this paper uses Item-Based algorithm,Ebinhaus forgetting function and Dirichly model for reference,proposes an agricultural information push method based on collaborative filtering algorithm,and designs a dynamic user interest model.The model consists of the data results processed by the algorithm and the corresponding users.Using the crawler technology,some news on agricultural websites are collected to the computer,and the downloaded news is processed in a series of ways,such as eliminating repetition.Based on user similarity in collaborative filtering algorithm,the improved hybrid user similarity is used to calculate user similarity,which can be divided into less browsing record stage and more browsing record stage.The user similarity can be calculated by collecting the browsing times of each user’s news,forming a set of neighbors among users,and calculating the similarity between news and news.Moreover,it predicts news scoring,ranking news scoring from high to low.Using weight calculation method,Ebinhaus forgetting function and Dirichlet model,a dynamic interest model of farmers is constructed.The model is tested on mobile phone.The test results show that Dirichlet model can accurately calculate users’ interest in news types,and users scoring in this model.It is higher than the ordinary push system,and can find out the user’s interest to a certain extent,and the mobile phone test shows that it can make the vast number of agricultural practitioners easy to use. |
