Page 81 - Contributed Paper Session (CPS) - Volume 5

CPS1060 Taik Guan T. et al.

Using machine learning technique to classify
geographic areas with socioeconomic potential
for broadband investment
1
1
2
Taik Guan Tan , Dino Isa , Yee Wan Wong , Mei Shin Oh
1
1 University of Nottingham, Malaysia Campus, Faculty of Engineering,
Selangor, Malaysia
2 CAD-IT Consultants (Asia) Pte Ltd. Selangor
Abstract
The telecommunication service providers (telco) commonly use the return on
investment (ROI) model for techno-economic analysis to strategize their
network investment plan in their intended markets. The numbers of
subscribers and average revenue per user (ARPU) are two dominant
contributions to a good ROI. Certain geographic areas, especially the rural
areas are lacking in both dominant factors and thus very often fall outside the
radar of telco’s investment plans. The government agencies, therefore,
shoulder the responsibility to provide broadband services in rural areas
through the implementation of national broadband initiatives and regulated
policies and funding for universal service provision. This paper outlines a
framework of machine learning technique which the telco and government
agencies can use to plan for broadband investments in Malaysia and other
countries. The framework is implemented in four stages: data collection,
machine learning, machine testing and machine application. In this framework,
a curve-fitting technique is applied to formulate an empirical model by using
prototyping data from the World Bank databank. The empirical model serves
as a fitness function for a genetic algorithm (GA) to generate large virtual
samples to train, validate and test the support vector machines (SVM). Real-
life field data from the Department of Statistics Malaysia (DOSM) for
geographic areas in Malaysia are then provided to the trained SVM to predict
which areas have the socioeconomic potential for broadband investment. By
using this technique as a policy tool, telco and government agencies will be
able to prioritize areas where broadband infrastructure can be implemented
using a government-industry partnership approach. Both public and private
parties can share the initial cost and collect future revenues appropriately as
the socioeconomic correlation coefficient improves.

Keywords
Socioeconomic; Broadband Investment; Machine Learning; Support Vector
Machine; Genetic Algorithm

1. Introduction
According to the World Bank’s report (2009), every 10% increase in
broadband penetration in developing countries will accelerate economic GDP

70 | I S I W S C 2 0 1 9