With the increasing amount of data and the growing need to transform it into accurate and useful knowledge, the Multi-relational Data Mining process becomes essential for knowledge discovery and decision-making. The extraction of association rules is a well-established and widely applied method for Data Mining that generates a collection of rules that relate two or more attributes. However, this process often generates a big amount of rules from which only a few are relevant to the analysis. This work proposes the use of Templates, pre-defined formats for the association rules, for the enhancement of results in Multi-relational Data Mining based on the generation of a result set with fewer rules and greater relevance. The validation of the technique will be given through the application in real and synthetic public databases and the application in environments for the management of real data relate to human health, such as Work accidents, epilepsy and hepatitis, and patients with mental pathologies and chemical dependencies.
News published in Agência FAPESP Newsletter about the scholarship: