Study Of The Agricultural Data Mining Platform Based On Classification And Clustering

With the widespread application of information technology,the agricultural industry stores a large amount of data,mainly including meteorological information,soil and soil quality information,information on diseases and insect pests,crop yields,irrigation and other information,and the amount of this information continues to increase.Data mining technology can find potential and useful knowledge information from the data,and can provide help to improve the quality of agricultural data information and solve agricultural production problems.The analysis and research of data is carried out on the basis of data acquisition and storage,and the graphical display of the analysis results can help users intuitively understand.To this end,building an agricultural big data platform that integrates data storage,analysis,and interaction can better manage and apply agricultural data.Big data application technologies such as Hadoop and Spark can perform large-scale data storage and rapid calculation.The integration of these technologies and website development technologies provides support for the realization of data mining platforms.This paper analyzes the situation of agricultural data mining and big data technology,selects the classification and clustering algorithm,develops a mining platform for agricultural data processing,collects and manages the information generated in the production process,and provides data mining methods for users to call.The main work of this paper includes: analysis of the application of data mining technology in agriculture,selection of classification and clustering mining algorithms for agricultural data,specific analysis of agricultural data using classification and clustering,and design of the architecture and function of agricultural data mining platform,And integrate Spark distributed memory computing framework,HDFS distributed file system,SSM(Spring + Spring MVC + Mybatis)website development and other technologies to build a highly available data mining platform.In the thesis,a detailed design of a data mining platform based on classification and clustering algorithms is carried out,mainly involving the overall architecture of the system,the key technologies adopted,the design of functional modules,the sorting of business processes,and database design.The architecture of the platform includes several levels: display layer,logic layer,calculation layer,and data layer.The data layer is divided into storage layer and collection layer,and it is composed of four sub-modules: data management,data preprocessing,data mining,and system management.On the basis of algorithm analysis and system design,a Spark cluster is built,Java language is used for development and a unified technology ecosystem,and some algorithms are implemented using Spark related components.After system testing,the developed platform can realize the functions of production process data supervision and storage,agricultural data knowledge base management and mining analysis processing,and provide assistance for agricultural production process traceability and smart production.