Research And Application On E-commerce Platform Intelligent Recommendation Technology Of Agricultural Products Based On Big Data

With the rise of big data processing technology and e-commerce platform for agricultural products,thus the species and channels of agricultural products market become more and more diversified.The sales data of agricultural products e-commerce platform is characterized by a large number of changes,however,as the agricultural products are not well received on the ecommerce platform,users can hardly pick up what they want at their first purchase.Therefore,an intelligent recommendation method of the agricultural products on the e-commerce platform has been an important measure to meet the individual needs in short responding time.Traditional recommendation methods are time-consuming and inefficient,so this paper proposes the research on the intelligent recommendation method of agricultural products based on big data processing technology,which can provide intelligent decision-making foundation for merchants in accurate sales and for users in personalized consumption on the e-commerce platform.The main research contents are as follows:(1)It proposed the application research of agricultural big data processing based on Spark platform.By comparing the performance of Hadoop and Spark data processing platforms,the results of Spark platform are faster and more accurate than those of Hadoop platform under the same large amount of data.The verification results show that the Spark platform is more suitable for the effective processing of agricultural big data,determine the Spark platform to process agricultural big data methods,and conduct further research.(2)It conducted the study on agricultural product intelligent recommendation method based on the theme weight fusion collaborative filtering algorithm.Firstly,a kind of LDA-MF(Linear Discriminant Analysis-Matrix Factorization)hybrid algorithm has been formed by integrating the document theme algorithm and matrix factorization algorithm.Secondly,a theme weight fusion collaborative filtering algorithm has been formed by weighting fusion of collaborative filtering algorithm based on the item and LDA-MF hybrid algorithm and the intelligent recommendation method for agricultural products has been verified.The experimental results show that the theme weight fusion collaborative filtering algorithm can describe the correlation degree between agricultural products and provide benign technical support for the similarity calculation of agricultural products.According to the three evaluation criteria of accuracy,diversity and RMSE,it is verified that the theme weight fusion collaborative filtering algorithm is advantageous than the single recommendation algorithm,which can provide reliable technical support for the intelligent recommendation system.(3)It designed and developed an intelligent recommendation system for agricultural products based on Spark platform.The above research results are integrated on the Spark streaming computing platform,and the Spark platform-based agricultural product intelligent recommendation system with four functional modules of agricultural product inquiry,intelligent recommendation,agricultural product management and user management has been constructed.The agricultural product inquiry module realizes the functions of user actions of data set processing;the intelligent recommendation module provides a variety of intelligent recommendation method verification functions;the agricultural product management module can realize the model construction of agricultural product evaluation theme and analysis function of sales trend dynamic;The user management module ensures complete inclusion and timely update of the user's agricultural product purchased behavior information.According to the modules,the system achieves an effective improvement in matching efficiency and accuracy of agricultural product sales and purchasing behavior,and provides a good basis for e-commerce platform intelligent recommendation of agricultural products.This paper has combined a variety of big data technologies to analyze the behavior characteristics of users on agricultural products' e-commerce platform,recommendation interesting products for users,expanded sales chances for merchants,which can greatly increase the probability of agricultural products sale,reduce their operating costs,and solve the problem of agricultural products transaction.Moreover,the solutions can provide strong support for the improvement of the national economy.