Granular Knowledge Discovery Framework ∗ A Case Study of Incident Data Reporting System
Adam Krasuski · 2013
A platform for fire & rescue incident data reporting system (IDRS) is presented as an example how the domain knowledge driven granule formation can assist in knowledge discovery and decision support. The current modeling, moni- toring and reporting systems rarely take advantage of semantic background of the analyzed phenomena. We discuss how to build and tune practically meaningful mod- els of processes by means of granules approximating their states and instances. We show how the layers of model creation should interact with lower-level layers of data preparation and transformation. We illustrate the proposed methodology by several IDRS related use cases. We also discuss the complexity of available data sources that can be utilized to make the proposed approach more useful.