A Hybrid Data Mining Model for Effective Citizen Relationship Management: A Case Study on Tehran Municipality

Currently, many governments are actively promoting implementation of ICT to be more citizen-oriented. For effective citizen relationship management, it is important to identify the needs of different citizen groups and to provide respective services for each group accordingly. In this way, the application of data mining tools would be very useful to understand citizen's needs. In this paper, focusing on the CiRM concept, we apply a data mining framework on the database of Tehran municipality. This framework consists of clustering and the association rule to improve citizen satisfaction. The main objective is to find the factors those affect the rate of satisfaction. Firstly, we use the K-means algorithm to cluster the subjects that cause citizens complaint. Every data point is identified in terms of the following features: the frequency, the number of days that at least one complaint occurred and the interval time between the first and the latest time of each subject during a season. Secondly, the association rule is used to identify the factors that affect the rate of satisfaction in the cluster of subjects that occur regularly during the season and have a high number of complaints. The results of the research are very useful to build a strategy recommendation system in order to improve the rate of citizens' satisfaction. This study could be notable as one of the first studies on using data mining tools in CiRM.