Title | ||
---|---|---|
Predicting Classification of Telecommunication Subscribers Based on Polytomous Logistic Regression Usage Model |
Abstract | ||
---|---|---|
Validating the performance of data mining models requires data not seen during model creation. In this paper, a bootstrapped partition of a full dataset has been used to gauge the ability of polytomous logistic regression models in predicting classifications of telecommunication subscribers. Coded coefficients of significant predictors facilitated the calculation of log of odds attributed to each group of users and formulation of logistic regression equations. Converted from the results are executed probabilities used for obtaining distinct group memberships. The classification rate obtained from cross-tabulated counts and percentages confirmed the capability of the model. With accurate predictions, the models could provide decision makers a better understanding of individual interactions and group-level behaviors of subscribers, a tipoff for improvement of Short Message Service (SMS) schemes to positively leverage service usage to generate more profit while extending customer satisfaction. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1145/3055635.3056615 | ICMLC |
Field | DocType | ISBN |
Data mining,Telecommunications,Computer science,Bootstrapping,Logistic model tree,Artificial intelligence,Odds,Logistic regression,Customer satisfaction,Pattern recognition,Multinomial logistic regression,Supervised learning,Polytomous Rasch model,Statistics,Machine learning | Conference | 978-1-4503-4817-1 |
Citations | PageRank | References |
0 | 0.34 | 0 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Rosmina Joy M. Cabauatan | 1 | 0 | 0.34 |
Bobby D. Gerardo | 2 | 27 | 13.79 |