JDM-based User Behavior Analysis On Mobile Internet

Mobile Internet was one of the most potential fields today. In the field, user behavior analysis was also becoming increasingly important. Developers were constantly developing and upgrading products in the process, but also needed to understand the user's needs.By collecting and analysing user behavior data, developers could further improve products and enhance the user experience. A large amount of theoretical research was did in the data mining and data analysis'field, but the amount which the research applied to production was limited. One of the main reasons was the mismatch between research and application. Most small and medium enterprises still remained using a simple SQL database and data statistics, which was the low efficiency of data reporting stage. The mining tools like SAS and SPSS had a high learning requires and expensive usage fees, while lack flexibility for some special algorithms and functions. There was a problem that how theoretical research of data mining be applied to industry.As some specificity of the mobile Internet products, the methods of analysis on the traditional Internet cannot be copied. The generation of JDM(Java Data Mining) was that developers wanted JDM to do data mining work in database just like "JDBC connects to database". But as JDM was not widely used, there was few related resources and cases. With a mobile Internet product, by the help of JDM API which provided a number of data mining algorithms and Oracle's OJDM(Oracle Java Data Mining) specific algorithm, the paper did the following work:got the original data of the client, stored it, extracted it, and then transformed it to the statistical analysis. In the process of studying and understanding JDM, the paper preliminarily explored some methods and ideas in the area of user behavior analysis on the mobile Internet.With the following aspects, the paper finally used JDM to do the work of analysing the user'behavior in the mobile product.1. It explored the cause of JDM and the architecture of JDM API, discussed the relationship between JDM and OJDM, gave code examples, and so on.2. It discussed the characteristics of mobile Internet products, the differences of data collection between the mobile Internet and the traditional Internet. With a specific product, it discussed in detail how the server collected user behavior data, how the data was extracted,transformed, and finally entered into the data warehouse.3.With JDM, the paper showed the process of mobile Internet user behavior analysis. First it briefly descripted the decision tree algorithms and the classical Apriori algorithm. Then combined with actual product data, it did the work of mining analysis from the user behavior data by JDM, gave the main code and obtained some main conclusions. Finally, the article listed other simple ways of data analysis.