The Design And Implementation Of Electronic Business Data Analysis Platform

The developing direction of computer application in the future is to strengthen the depth of the data processing.After the concept of big data is put forward,data's acquisition,processing,mining and application in the study have become a hot research in various industries.Electronic commerce(shorted as e-commerce)has changed the traditional business model.It is one of the most successful applications of the Internet.And a large amount of data generated in e-commerce also has very important application value.At present in the field of e-commerce,data analysis and its application are not very extensive.It usually pays attention to one-particular-aspect research analysis,and can not conduct a comprehensive analysis of the data.On the basis of such shortcomings,the thesis proposes the e-commerce data analysis platform model.The model mainly includes five parts which are the data source layer,operative data layer,data integration layer,data processing layer,and data application layer.The data source layer is composed of a plurality of heterogeneous database,which store different types of data operated in the process of e-commerce.2.In the overall framework,operative data layer extractes mainly "hot" data in some period from data source layer and realizes the data inquiry.According to the requirements of application,data integration layer integrates the data from operative data layer to provide the foundation for data application.Data processing layer adopts the methods of data mining,online analytical processing and data statistics to process the integrated data.Data application layer is based on the users' business needs to construct data processing and application,to call the various methods of data analysis,and finally to get the analysis results.In order to give e-commerce data analysis platform the ability to deal with massive data,the platform introduces Hadoop technology in specific implementation.On the basis of HDFS distributed storage model,the platform improves not only the ability of storage and processing of unstructured data,but also the expansion capability of the system.3.Aiming at the processing efficiency of massive data,the platform is combined with Map/Reduce parallel computing technology as the tool to process the mining data which improves Apriori algorithm and operative efficiency of classified data mining.The model designs the framework of application layer in e-commerce data analysis platform.Through the discussion of business process of e-commerce data analysis,the model also determines the application goal,discusses system fuction design and realization process,and finally checks its effectiveness through the tests.E-commerce data analysis platform adopts the hierarchical structure model which makes full use of intelligent data processing technology to analyze the data from the platform in order to meet different application needs.It improves the business model for e-commerce system managements.It helps to maintain the relationships with customers better.It has strong practical value to provide effective tools to improve economic efficiency.It also has theoretical research value on the basis of study and improvement of data mining algorithm.