Construction And Research Of Agricultural Planting Information Recommendation System Based On Hadoop

As the highest agricultural output value in China,the planting industry plays an important role in the entire agricultural production and development.For the websites that provide agricultural information in China,there are certain problems such as redundant information,poor professionalism,and inaccurate information provision.Therefore,building a recommendation system for agricultural planting information and improving users' accuracy in obtaining planting information are helpful.Farmers' access to information and the production efficiency of the plantation industry have improved.This paper constructs an agricultural planting information recommendation system through Hadoop and other related technologies.By analyzing the technologies and methods adopted by the system,the overall architecture of the system is designed.The construction process of the recommendation system is mainly based on the recommendation module and the search module.The recommendation module is the core module of the system.The construction of the recommendation module mainly uses the item-based collaborative filtering algorithm provided by the machine learning framework Mahout.This paper through the construction and research of agricultural planting information recommendation system,mainly includes the following research work:Through the construction and research of agricultural planting information system,this paper mainly includes the following research work:1.Through the analysis of system requirements and functions,and using components provided by tools such as Hadoop,the establishment and deployment of the experimental research environment for the recommendation system was completed.2.By analyzing the recommendation algorithm in Mahout,the information recommendation of the collaborative filtering algorithm based on items is realized,and a non-linear relationship interest calculation method is proposed in the process.In view of the strong regional and time characteristics of agricultural planting information,a method to solve the cold start of the recommendation system was designed.The HBase filter was mainly used to provide users with region and time-related planting information,therefore,it effectively improves the recommendation efficiency in the initial stage of the system.3.As a supplement to the information channel of the recommendation system,using Solr search engine tools,users can quickly search the stored information in the system.In the cluster environment,it mainly constructs and studies the secondary retrieval method ofHBase.4.Through testing,the system can implement various functions of response and service in a cluster environment.And through the selection of the similarity calculation method and comparing the evaluation data of different environmental conditions,the results show that the construction of the system can improve the recommendation efficiency of planting information.