Research And Implementation Of Agricultural Planting Information Recommendation System Based On Network Embedding

The Internet has been an indispensable part of people’s lives.However,the information overload makes it impossible for users to obtain the information they need timely and quickly.Meanwhile,the emergence of recommendation system has become an effective way to deal with this problem.With the implementation of the national information strategy,much attention has been paid to the application of the Internet and its related technologies in the field of agricultural science and technology information services.As a crucial part of this field,how to transmit agricultural planting information to farmers accurately and quickly has become a research hotspot in the construction of agricultural informatization.Since the traditional recommendation algorithm underperforms in agricultural planting information recommendation,this kind of algorithm will face serious scalability problem with the expansion of data scale.The network representation learning technology provides a new idea for the research of the recommendation algorithm,which makes it possible to break the limitations of traditional recommendation algorithm.This paper is devoted to providing farmers with convenient services,and makes an in-depth research on the recommendation algorithm of agricultural planting information by utilizing the network representation learning technology.The main contents of this paper are as follows:(1)Aiming at the scenario of crop planting recommendation,this paper proposes a recommendation algorithm: Farm-Crop Graph based Network Embedding Recommendation.In this algorithm,the attribute information of farm and crop are considered comprehensively.Based on these information,the Farm-Crop Graph is constructed and the edge weight is quantified reasonably.Then it applies static network embedding methods to obtain the vector representation of the two kinds of nodes in low-dimensional space,and then mines the correlation between farms and crops.Finally it recommends crops through the correlation.Empirical experiments conducted on the Crop Plant dataset illustrate the efficiency of the proposed algorithms.(2)Aiming at the scenario of agricultural planting information recommendation,this paper proposes a recommendation algorithm: Network Embedding based Agricultural Planting Information Recommendation.In this algorithm,the information of user attribute,user behavior strategy,document attribute and time are considered comprehensively.Based on these information,the user relation network is constructed and the edge weight is quantified reasonably.Then it applies dynamic network representation learning methods to obtain the user feature vectors at different times.Finally it recommends information through the similarity between users.As the contribution of different context information to similarity calculation is different,the attention mechanism is introduced to learn the relevant weight coefficients.Verified by the Agric Plant dataset,this algorithm outperforms other comparison algorithms.(3)Based on the above two recommendation algorithms,the agricultural planting information recommendation system is implemented by HTML5,Django and other related technologies in this paper,which includes crop planting recommendation,agricultural planting information recommendation and other functional modules.Users will enjoy a better agricultural information query experience through the simple and fast interactive page.