One small step for a man
One Giant leap for the mankind

There is no wealth like Knowledge
                            No Poverty like Ignorance
Journal of Emerging Trends in Computing and Information Sciences Logo

Journal of Emerging Trends in Computing and Information Sciences >> Call for Papers Vol. 8 No. 3, March 2017

Journal of Emerging Trends in Computing and Information Sciences

Value Based Time Dimensioned Churn Prediction

Full Text Pdf Pdf
Author Sana Salman
ISSN 2079-8407
On Pages 180-183
Volume No. 4
Issue No. 2
Issue Date March 01, 2013
Publishing Date March 01, 2013
Keywords ADS, Prepaid Churn, ARPU, Value based segments, Residual, Outlier, Leverage


Abstract

The advent of de-regulation in year 2004 and induction of new market players resulted in growing competitive telecommunication industry in Pakistan. Subscriber retention is a vital focus area for the operators. Churn management strategies have gained importance in the current age of hundreds of competitive offers attracting subscribers to move from one operator to another on the fly. Data mining tools help to build strategic infrastructures so that predictive and descriptive statistics could be used well before time to churn. [7] In this paper we present a comparison of real time event triggered churn models to conventional monthly churn models in which the problem arises when the event triggered behavior of the subscriber is averaged out. It is an iterative process of predicting churn based on analytical data set assembled in different time dimensions. Logistic regression is used as the primary classification technique. The data exploration moves through the usage based behavior of the subscriber. The transformed analytical data set results in capturing the trend based behavior of the subscriber in a much smarter fashion than the base data set. Time dimensional approach gets its real flavor because of the hourly update of the subscriber activity in the data ware house. The modeling and scoring is done on real data as part of the implementation. The end result is a score card that assigns an average score to the subscriber’s propensity to churn based on all models. Our experimental results indicate that in order to capture the churn happening any time in the life cycle of a subscriber I-e from activation to maturity, the multi-dimensional time based analytical approach is a viable solution.
Back

Seperator
    Journal of Computing | Call for Papers (CFP) | Journal Blog | Journal of Systems and Software | ARPN Journal of Science and Technology | International Journal of Health and Medical Sciences | International Journal of Economics, Finance and Management     
Copyrights
© 2015 Journal of Computing