A Study On Data Prepare Of Home Client In Telecom Indurtry Based On Rough Set And Cloud Model

Telecom industry in china is a fast growing industry, but also facing intense competition. Client is the key resource of telecom companies because of the trend that mobile replacing of telephone, especially the VIP client. Knowing more about your client is the key to get and reserve client. china telecom jiangxi branch company want to know more about their home client according behavior-class attribute, financial-class attribute, background-class attribute, cost-class attribute, but there are more 500 attributes included in this four classes attribute. These papers try to prepare data form china telecom Jiangxi Branch Company based on rough set and cloud model, and select these key attributes for classifying home client.This papers presents and analyze some kinds of data preparation steps, then, suggests a set of steps of data preparation: data cleaning, data integration, data transformation, data reduction and introduce how to do in every step.The frame of this papers is clustering the 23 attributes suggested by experters in telecom indurstry; the outcome of clustering is the standard of the clustering ability of data after preparation. So this papers firstly completing the missing data according the missing reason, second, cleaning out the outliers using SAS.This papers discretize data from Jiangxi Branch Company using cloud model, compared to other redundancy, cloud model has a better performance because of its uncertainty. This papers has analyzed the source of redundancy, suggested an attribute selecting redundancy according the importance of attribute using rough set as a measurement. Make a comparison between before attribute selecting and after attribute selecting after having finished the selection process.Finally, there is a summary and outlook of this papers.