Knowledge Discovery: Data Mining by Self-organizing Maps

Abstract

Due to the characteristics offered by automated management systems, municipal administrations are now attempting to store digital information instead of keeping their physical documents. One consequence of such fact is the generation of large volume of data. Usually, these data are collected by ICT technologies and then stored in transactional databases. In this environment, collected data might have complex internal relationships. This may be an issue to identify patterns and behaviors. Many institutions use data mining techniques for recognize hidden patterns and behaviors in their operational data. These patterns can assist to future activities planning and provide better management to financial resources. Intelligent analysis can be realized using the Support Tools and Support Decision Making (STSDM). These tools can analyze large volume of data through previously established rules. These rules are presented for STSDM in the training phase, and the tool learns about the patterns that should look. This paper proposes a model to support decision making based on self-organized maps. This model, applied to electronic government tools, can recognize patterns in large volume of data without the set of rules for training. To perform our case study, we use data provided by the city of Campinas, Sao Paulo.