| Summary: | 碩士 === 輔仁大學 === 應用統計學研究所 === 94 === Based on the characteristics that the local cable TV (CATV) owns an individual business area and original cable viewers and the advantages of the hybrid fiber/coaxial (HFC) and Broadband network skills, CATV can dig out the potential customers of the Broadband network and develop new customers effectively by applying the theories of data mining and customer relationship management (CRM), etc. This research can be used for further research by academics or the Broadband Cable industry’s reference.
This research is based on the current cable TV audiences (about 2,000 data) that were surveyed by Chi-square test and the process of data cleaning in order to check the correlation between each variable. During this process, the correct data variables will be culled. Lastly this research will find the best model of the customer prediction by decision tree technique, and build an excellent customers response model for the Broadband network industry. The objective of this research is to enable the industry’s output value and aim at the group of best customers, and then promote effectively. On the other hand, it will help decrease the cost of the promotion but increase the life time value of customers.
The data mining technique that we apply in this investigation actually gets positive results of prediction. However the accuracy of the anticipation model is affected due to the limited amount of data in this research. As the market of Broadband cable is developing, the database will keep growing and the capacity of this model will also grow as well.
|
|---|
