Realization And Data Analysis Algorithm Of Agricultural Big Data Platform

With the development of agricultural informatization and the accumulation of data with respect to crop framing,animal husbandry,fishery,processing of agricultural products and meteorology,agricultural data is growing at an unprecedented speed.Agricultural data in China is characterized by real-time,high dimension,dispersed storage,and difficult to analyze.For one thing,agricultural structure in China is complex,agricultural data involves various fields.For another,agricultural data are vulnerable to geographical environment,soil,weather,crops,pests and so on.To utilized these data requires the support of big data platform.The big data platform can integrate the data in the field of agriculture,providing query,download,upload,visualization and other functions.The data mining method based on the agricultural big data platform can reveal the knowledge behind in the agricultural data and discover useful rules.Big data platform can also make advice and guide the agricultural workers.Therefore,it is of great significance to develop agricultural data platform with the above functions.According to the properties of agricultural big data,this work focus on data analysis and visualization and proposed a spectral improved clustering algorithm.This work also built an agricultural big data platform with data mining function.The main work is as follows: We construct an agricultural database by collecting data from China statistical yearbook,experimental data in the university,at the same time,we also grab a lot of agriculture data from related website using web crawler technology.In the research of the key technologies of agricultural data platform,we propose an improved spectral clustering method for data mining.The result of the simulation experiment of seeds and soybean data sets in UCI data set shows that both the accuracy and speed are improved.We use multidimensional data visualization technology to realize the data visualization function.In the platform development,we give the system design scheme and realize all the functions of the platform according to the platform requirements using J2 EE technology.In the aspect of system testing,we use the agricultural machinery,animal husbandry data to carry on the experimental simulation.The agricultural big data platform proposed in this dissertation,have a user-friendly interface and one can use the system to upload,download or query the agricultural data or use it for data mining and visualization.In the aspect of data collection,the network crawler module can update the system database automatically.The research on the key technology of agriculture big data and the develop of agriculture big data platform has important valuation in agricultural informatization and intelligent agriculture.