The Data Problem in Data Mining

International audienceComputer science is essentially an applied or engineering science , creating tools. In Data Mining, those tools are supposed to help humans understand large amounts of data, and produce actionable insight. In this talk, I argue that for all the progress that has been made in Data Mining, in particular Pattern Mining, we are lacking understanding of key aspects of the performance and results of pattern mining algorithms. I will focus particularly on the difficulty of deriving actionable knowledge from patterns. I trace the lack of progress regarding those questions to a lack of data with varying, controlled properties, and argue that we will need to make a science of digital data generation, and use it to develop guidance to data practitioners