Research On Agricultural News Recommendation System Based On Collaborative Filtering

Internet technology provided new impetus to the development of agricultural modernization,and endowed producers and consumers with unprecedented information capabilities,but it has also brought a huge amount of information which makes it difficult for people to sort out the useful information in the vast ocean of information and wastes a lot of resources.Therefore,it is an urgent need to find a way that can automatically help people identify and provide what they need.Information recommendation,also called recommendation system is an option.This paper focused on the collaborative filtering recommendation algorithm and its application in agricultural news recommendation,and designed and implemented an agricultural news recommendation system based on collaborative filtering.The main contents of this study are as follows:(1)The recommendation system and its technology are firstly introduced,then four recommendation algorithms including content-based recommendation algorithm and association rules recommendation algorithm,and user-based recommendation algorithm,the item-based recommendation algorithm are illustrated in detail;(2)Collaborative filtering recommendation algorithm was the research priority,and two groups of comparative experiments were conducted based on agricultural news and Movielens data set.In the first experiment,the four recommendation algorithms were compared with news data.The results of the experiment showed that the item-based recommendation algorithm was better than the user-based recommendation algorithm in terms of MAE and RMSE,and the latter was better than the content-based recommendation algorithm and the association rule-based recommendation algorithm in terms of precision and recall rate.In the second experiment,a comparative analysis was made between the user-based recommendation algorithm and the item-based recommendation algorithm.It is consistent with the results of the first experiment,namely,the user-based recommendation algorithm was better than the item-based recommendation algorithm in terms of precision and recall rate,while the latter was better than the former in terms of mean absolute error(MAE)and root mean square error(RMSE);(3)The agricultural news recommendation system was designed and implemented based on the performance of collaborative filtering recommendation Algorithm in the previous experiment,Firstly,the overall architecture of the system was defined,and the three-tier structure was adopted.Then the functional modules and database were designed.Finally,the system was implemented.