When using data mining to find regularities in data, the obtained results (or patterns) need to be interpreted and understood, so that they can be subsequently reused in new Knowledge Discovery processes. The background knowledge to interpret these patterns is, however, not always straightforward to find, as it might require different sources and domains. This intensive process is usually committed to the experts in the domain, whose work is to analyse the results of a data mining process, give them a meaning and refine them. With the rise of Linked Data and the increasing number of connected datasets, we assume that the access to this background knowledge can be easier, faster and more automated. In this talk, I am going to present my PhD research, aiming at demonstrating how Linked Data can be used to provide the background knowledge, reducing the human effort put into the results interpretation.

Watch the webcast replay >>