عنوان مقاله [English]
Business intelligence refers to computer-aided techniques for the discovery and analysis of business data. It aims at supporting better business decision making. At the present time, with the development of information systems, organizations are moving toward an optimum use of information, and expanding information to applicable knowledge in order to achieve more business intelligence. Business intelligence and customer relationship management systems are playing a critical role in the creation and development of value
added techniques for organizations. As a result, municipalities, as the main organizations in city management, are moving towards the development of a powerful information technology infrastructure. They provide an appropriate business intelligence research area, due to their enthusiasm for development, their access to a large number of customers and the complexity of their relationship with them, i.e. citizens. Among all revenue sources, the imposition and reconstruction section has attracted municipalities, such as that of Esfahan, as a source of more capital development. This research aims at determination of the hidden behavioral patterns of citizens in the imposition and reconstruction section by means of data mining methods. In order to achieve this, data related to landowner payment status, which is land, landowner specifications and their yearly payment details for the last seven years, are analyzed. The main objective of this research is to understand factors affecting citizens being good or bad, regarding payment of impositions.
Accordingly, in this case study, initial steps for the achievement of business intelligence are analyzed using the K-Mean algorithm. K-Mean is a clustering method for grouping similar objects in the same group. Considering various specifications of land and land owner, citizens are clustered into five account sets of: very good, good, half good, bad and very bad, based on payment delay.
Consequently, most effective factors affecting citizen imposition payment behavior are determined for future predictions. This is achieved by means of neural network and decision tree methods. The results of the two approaches were not exactly similar; however, both suggested two similar factors as the most influential in predicting customer payment behavior. These two factors can be considered in a further customer relationship management plan, with respect to higher business values over time.