|
Title:
|
A FRAMEWORK FOR PREDICTIVE DATA MINING IN THE TELECOMMUNICATIONS SECTOR |
|
Author(s):
|
Adrian Costea , Tomas Eklund , Jonas Karlsson |
|
ISBN:
|
972-9027-53-6 |
|
Editors:
|
Pedro Isaías |
|
Year:
|
2002 |
|
Edition:
|
Single |
|
Type:
|
Full Paper |
|
First Page:
|
38 |
|
Last Page:
|
46 |
|
Language:
|
English |
|
Cover:
|
|
|
Full Contents:
|
click to dowload 
|
|
Paper Abstract:
|
In this paper we construct a framework that enables us to make class predictions about telecommunication companies financial performances. We have two goals: to validate our methodology and, using it, to gain insights in this relatively new and very sensitive industry: the telecommunications sector. We have obtained high accuracy rates for the classification models, and small differences between training and test dataset accuracy rates. The two classification techniques have performed similarly in terms of accuracy rates (decision tree, slightly better than logistic regression) and class predictions (all companies selected were placed in the same clusters by both methods). We have analyzed the movements of the largest four telecommunications companies. The results show a strong connectivity with what had really happened to these telecommunication companies during the second part of the last decade, and the beginning of the current one.
KEYWORDS :
SOM algorithm, class predictions, telecommunication sector. |
|
|
|
|
|
|
