The goal of the doctorate project is the creation of an association rule mining (ARM) framework that combines data mining techniques and a domain ontology. The framework aims to automate and improve the post-processing stage in ARM. The system will be fed with the association rules generated, from a specific database, by a traditional ARM algorithm. The output will be a small set of potentially more useful and interesting ARs. The system will contribute to reductions in time and cost of tasks executed by domain specialists in AR post-processing; with the increase in the quality of the data mining process results, given that it will present a reduced and well-selected amount of information; and with the expansion of the potential for delivering useful knowledge of ARM algorithms.

The proposed approach will use the data from the Hospital Cancer Registries (RHC - acronym in Portuguese) and might be combined with data from other sources such as the National Registry of Health Establishments, the Brazilian Institute of Geography and Statistics and the National Supplementary Health Agency (ANS). And this integration of databases from different sources represents an important step in extracting interesting and non-obvious associative patterns.