Integration of geographical information systems and artificial intelligence technics to support banking industry decision taking

Abstract

Making a decision on where would be the best location to sell a product is the most
crucial task in any market sector. However, this is a complex undertaking since it
involves numerous variables, not only related to space, but also related to
subjectivity and symbolism and often related to personal opinions. Taking into
consideration that the decision on the location involves many variables as to space
as well as symbolic variables, this paper presents the integration of a Geographical
Information System (GIS) with an Expert System (ES) in order to help with the
decision taking process of where to start a new retailer in the banking sector, in this
case within the HSBC Bank Brazil S/A, Multiple Banking, Department of Real
Estate Administration for the area of the city of Curitiba, Paraná. The system was
developed in several steps. To implement GIS ArcGIS was used and to develop the
Expert System it was used a shell, named Expert SINTA. The user interface was
developed by using programming language Visual Basic. A comparison was made
between the methodology used of the specialists of the Department of Real Estate
Administration and the proposed system in order to validate the results. The
conclusion was that the integration of the systems contributed to an improvement
within the decision maping process as to the location in the banking sector,
enlarging the possibilities of the spatial analysis and helping the experts with the
tasks that involve subjective estimations and often opinions.