The Design And Implementation Of Recommendation System For Agricultural Trading

To meet the needs of consumers to obtain personalized information, recomendation system has been rapidly developed, and it has been widely applied to various fields, particularly in the area of electronic commerce. In the e-commerce system, the system provides consumers the goods, which are associated or they may interested in, and it also can provide personalized information services for the needs of different consumers, what's more it is convenience for consumers to obtain information, at the same time it also enhance the loyalty of users. In the background of "Internet +", the rural information process has begun to advanced gradually. "The new rural community of e-commerce and logistics information service system" as one of the core building of Henan province major projects(integration and application the key technology of new rural community information service) which is facing all kinds of agriculture-related information, use the recommendation of algorithm as the core support, according to supply and demand information to match, to provide consumers with intelligent information matching services?At the same time it can promote the circulation of agricultural products, accelerate the industrialization process of agriculture and increase farmers' income, which has the important practical significance.Based on articles collaborative filtering algorithm is widely used in the electricity supplier system.The algorithm is used to evaluating the similarity between the articles of different items by user ratings,and make recommendations based on the similarity between the goods. Moreover, the algorithm has many advantages such as it needn't have the domain knowledge, personalized, a high degree of automation and as time goes on, it can self-improve performance. However, in the practical application of the algorithm,we should make some adjustments or improvements combine with the specific application. In particular: itcan provide a flexible way based on taking full account of the association relationship between the goods.In the aspect of computing the association relationship: first, create co-occurrence matrix items to describe the relationship between any two objects; then construct items scoring matrix, which is used to characterize the degree of correlation between the articles, provide data support for the similarity calculation between the articles. In the aspect of the way of recommendation: this paper puts forward two ways: price priority and distance priority. When elect different ways to recommend, respectively consider the price attributes and geographical attributes to build different recommendation rules to make the system flexible and stable personalized recommendations.The collaborative filtering recommendation algorithm based on goods is one of the widely used algorithms in the electricity supplier system, which is used to evaluating the similarity between the articles of different items by user ratings, and make recommendations based on the similarity between the goods.Moreover, the algorithm has many advantages such as it does not need to have the domain knowledge,personalized, a high degree of automation and as time goes on, it can self-improve performance. However,in the practical application of the algorithm, it should combine with the specific application to make adjustments or improvements. In particular: in this paper, on the basis of taking full account of the association relationship between the goods, provides a flexible way. In the aspect of computing the association relationship: At first, create co-occurrence matrix items to describe the relationship between any two objects; then constructing items scoring matrix, which is used to characterize the degree of correlation between the articles, provide data support for the similarity calculation between the articles. In the aspect of the way of recommendation: this paper puts forward two ways: price priority and distance priority. When electing different ways to recommend, this paper will respectively consider the priceattributes and geographical attributes to build different recommendation rules to make the system flexible and stable personalized recommendations.Based on the analysis of actual needs and technology roadmap of new rural community e-commerce and logistics information service system, design and implementation of oriented agricultural trading system are carried out on the basis of the core algorithm which is based on articles collaborative filtering algorithm. The main work is as follows:Firstly, System requirements analysis. First, analyze the overall demand for the system. The article is about the object model and system data flow, based on the over all demand, refine requirements of the recommended system from its business needs and data requirements.Secondly, Architecture design of the recommended system. According to the actual needs of the recommendation system, we design a stable and efficient architecture for the recommended system. Take the recommended system's business needs and data requirements as the starting point, design the data model reasonably. And analyze the items of collaborative filtering algorithm on the premise of the designed data model.Thirdly, design and implementation of the recommended system. Create article co-occurrence matrix to describe the relationship between any two articles. Construct materials scoring matrix, analyze score of different objects from different users to evaluate similarities among them, and calculate the right value for the users to recommend items. Different recommendation rules can generate different recommendation algorithms. In order to the upper layer of the caller to process data uniformly, we use the strategy pattern design ideas to encapsulate.Finally, verify the recommended system according to the data of the system.