Study On Web-based Agricultural Data Mining Platform

With rapid accumulation and growth of agricultural data, how to get scientific agricultural knowledge, rules and decision-support information by making use of the vast amounts of agricultural data, becomes very important. Data mining technology means to extract potential useful information and knowledge which is novel and unknown to people, from data collection. At present, data mining technology started late, it is used only when classical mining algorithm is applied in agricultural data processing, not to mention a set of systematic research and development methods. To solve the problems above, this paper presents a theoretic framework for Web-based agricultural data mining platform, including data mining visualization, data mining process and data storage, etc., and constructs a Web-based agricultural data mining platform on Responsibility Chain Pattern. At last, by some specific agricultural data application case, scalability, flexibility and effectiveness of Web-based agricultural data mining platform are verified. Major research work and contributions are as following:(1) It put forwards a multi-stage model on data mining and does deep research in data preparation, data mining and result evaluation, so to clearly define general agricultural data mining process and application features of classical data mining algorithms.(2) It put forwards a theoretic framework for Web-based agricultural data mining platform, and analyzes the key technologies as visualization of data mining, data mining process and data storage. The theoretic framework has the following main characteristics:improving platform interactivity by Web and visualization technology; making data mining algorithms to be dynamic substitution by configurable data mining process; bridging background multiple databases and data warehouse by unitary and scalable data integration, and so on. The theoretic framework helps to provide theoretical guidance to build an agricultural data mining platform and an expert system.(3) It devises a Web-based agricultural data mining platform by Responsibility Chain Pattern, including data preparation chain, data mining chain, data bridging chain and data filtering chain. The platform is successfully applied in data analysis as plant diseases and insect pests as well as agricultural soil pollution, so to verify not only effectiveness of theoretic framework for Web-based agricultural data mining platform, but also flexibility and scalability of Responsibility Chain Pattern.