چكيده به لاتين
Todays, with the advancement of science and technology and the development of organizations, the volume of stored data is increasing dramatically, so searching for these data and getting important and practical results becomes more difficult and the need for scientific knowledge to extract applied knowledge of data is felt. . Data Mining is a science that by searching in big data sources, discovers patterns and rules that simple statistical analyses are unable to do. One of the areas that requires using this tool to analyze large data and predict models, rules and patterns is the urban service area.
In the present study, for the data of years 1393 to 1395 (Iranian calendar) of the urban management system “137” using data mining knowledge, the information obtained from the request of citizens for addressing urban problems has been investigated. In this research, the clustering of 22 areas of Tehran's municipality based on three perspectives including (1) the average time interval between the occurrences of the issues, (2) the number of citizens' calls around the subject, and (3) citizens per call around the subject, is done. The optimal number of clusters are 2, 3 and 2, respectively. The survey of data of the city management system of Tehran municipality 137 based on the three perspectives is one of the contributions of this research. In the following, to predict the message status based on (1) region, (2) subject, (3) the implementation unit, and (4) the season, for each cluster obtained, the combined method derived from the Bayesian network methods, Neural network, logistic regression, as well as C5.0, C & R Tree and CHAID algorithms, decision tree with acceptable error of 25%. Determining the combination method for each cluster is another contribution of this research. Each of the outcomes of this study will help the managers and decision makers of Tehran municipality to predict the final status of registered messages in each of the 22 districts of the Tehran municipality based on the three perspectives.