Comparative Study on Decision Tree Techniques for Mobile Call Detail Record

Every mobile operator of today＇s world switches their technology over from 2G （second generation） to 3G （third generation） network. Operators are keen analyzing their CDR （call detail record） obtained over the past usage for predicting the behavior of their customers and their usage. The operators are willing to mine knowledge from real-world dataset which implies the pattern of user mentality on this changing world. To identify the usage of 2G and 3G services the classification models were trained using the data collected from PAKDD 2006 dataset. In order to obtain the prediction accuracy, the classifiers were evaluated using 10 folds cross validation. On comparing the results of the experiment, J48 performed more accurately and random tree consumed less time.