Research On Recommendation Models Of Agricultural E-commerce

E-commerce has become a shopping patterns that can be compared with entitiesshopping, It involves a very broad field like basic necessities of life，culture，entertainment and so on, you can say "pervasive", it brings convenient greatly forconsumers,while its side effects can not be ignored also. Agricultural e-commercesite, for example, agricultural commodity information, consumers demandinformation and a variety of other over-abundance of information on the Internet isfull of this highway, People’s needs and supply of electricity supplier should be liketwo car towards the same aim in the same highway, they get their desired bycollision, However, due to the proliferation of information, people’s needs andsupply of electricity supplier came to a different highway, they meet the needs ofeach other only when a traffic accident happen, and we know, it is unpredictable.Therefore, agricultural intelligent recommendation technology appeared. It put thecharacteristic of the rural market，farmers user’s characteristics and e-commerceproduct recommendation technology features together, straighten the path betweenthe supply and demand of electricity providers and consumers, so that they are met.The purpose of the text is to provide a good product for the agricultural user, wedesigned the model of agricultural intelligent recommendation propose threerecommended ways to improve the user’s shopping experience and therecommendations ability of agricultural products by analyzing the characteristics ofrural markets, farmers and agricultural product users.We found that attribute of project is a key of judging whether the user purchaseor not. Therefore, we determined the value of the property to the user and summedthem to get the true score by analyzing the user’s Ratings, and makerecommendations more scientific.And we know that the project-based recommendation algorithm are flawed,they can not handle the problem of new users, so we design recommendations basedon the user’s mode as a complement of project-based recommendation algorithm,and enrich the processing capabilities of the algorithm by Hadoop technology, make it have the ability which can handle complex issues like large data problems in future,enhance the recommendation system scalability.After completion the two models, we found there is still a problem that it cannot handle new product, so this paper designs keyword search recommendationmodel. This model found keyword by semantic network, then put them into differentclasses by TD-IDF Technology, and then decide to use what recommend algorithm.This model makes the search more efficient and avoid a lot of unnecessary problems.