Design And Research Of A Trading Platform For Leisure Agriculture Based On Data Mining

Leisure agriculture is an important industry of the people’s livelihood and new consumptionpatterns. It can increase efficiency of agriculture and the income of the farmers， it also canimprove the rural environment and made positive contributions to the development of economicand social. This paper uses the data mining technology and the theory of wireless sensornetwork， combined with the related technology of agriculture， research and analysis on thecollected environment parameters. Then design a new mode of leisure agriculture which is basedon modern computer technology， it integrating virtual and reality. Users can manage the realfarm via the Internet. At present， there are a lot of problems in the process of cultivation such asover-reliance on planting expert，lack of quantitative plant guidance and labor intensity. In orderto solve the problem， design the optimization model of Greenhouse control parameter， it canreplace the planting expert and do plant guidance for crop growth. The main work is shown asfollows:(1) The study on the optimal parameter extraction mechanism of crop growth process.Collect the environment parameters of crop growth which is planted by the experienced farmers，establish expert database to guide the production. Using the scoring system of evaluating theplanting effect to determine whether to update the optimal parameters of the expert database，ensure that the saved parameters is the optimal parameters， also can be changed next time. Inthe process it can save the experienced farmers’ wisdom. Then the system can complete the taskbetter even without the experienced farmers.(2) The study on the optimization model of Greenhouse control parameter. In this paper，the OCA objective cluster is applied to RBF neural network. It can solve shortcomings about thatthe RBF network cannot be objectively selected basis function center and improve the networktraining speed and accuracy rate. The improved neural network is introduced into theoptimization model of Greenhouse control parameter. Use the parallel distributed processingstructure of the neural network to complete the simulation expert reasoning process， achieveprecise management of greenhouse，and play an important role in the development of intelligentagriculture.(3) Design of commodity recommendation model. Based on the analysis of association rules mining algorithms， improve the large number of candidate sets produced dy the classic Apriorialgorithm. The green mall of trading platform looked as the application background， using theimproved Apriori algorithm to design the recommendation model. The model is mainly used toanalyze the user item order and dig out the hidden relationship between the sales of goods. Themining results are recommended to the user as ‘Hot commodity’， so it can improve the user’spurchase satisfaction and the sales of goods.The running results show that， the whole set of trading platform for leisure agriculture isdesigned by MVC development pattern of J2EE platform. It not only can reduce the coupling ofthe system software， but also can improved the reusability and maintainability of the system. Inthis paper， the optimal parameter extraction mechanism based on the updated design canautomatically update the optimal parameters of the expert panel， and more efficient. Thetraining speed and accuracy are greatly improved after the RBF neural network improved by theOCA objective clustering， and the optimization model of Greenhouse control parameterdesigned by the improved neural network can provide users with accurate planting guidance. Therecommendation model designed by the improved Apriori algorithm had improved obviously onthe execution time.