Ordinary Least Squares Regression Method Approach for Site Selection of Automated Teller Machines (ATMs)


Bilginol K., Denli H. H. , Şeker D. Z.

Spatial Statistics Conference, Avignon, France, 9 - 12 June 2015, vol.26, pp.66-69 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 26
  • Doi Number: 10.1016/j.proenv.2015.05.026
  • City: Avignon
  • Country: France
  • Page Numbers: pp.66-69

Abstract

The importance of alternative distribution channels on finance sector is increasing day by day. Most of the customers consider about ATM accessibility. It is an important criterion for customers' bank choices. Banks are making investments for increasing the number of ATMs to serve more customers and generate ease-of-use. Here, it is important to make the investments to appropriate locations for new ATM establishments. The ATM location distribution must be balanced while ensuring efficiency and answering demand of the costumers. On the site selection problems there are many spatial statistical methods such as Factor analysis, cluster analysis, neural networks, regression analysis and correlation analysis. In this study, canonical correlation analysis and ordinary least squares regression analysis are used. First of all, to find which criteria affect ATM site selection decision, all inputs are analysed with correlation method. Then, the result criteria which affect ATM locations are analysed on ordinary least squares regression model. Finally the optimum ATM locations and the predictive efficiencies of those ATMs are found. This study aims to make predictive analysis and to find optimum locations for ATMs by using ordinary least squares regression method. (C) 2015 The Authors. Published by Elsevier B.V