HiPaR: Hierarchical Pattern-Aided Regression

Luis Galárraga, Olivier Pelgrin, Alexandre Termier

Abstract: We introduce HiPaR, a novel pattern-aided regression method for data with both categorical and numerical attributes. HiPaR mines hybrid rules of the form p⇒y=f(X) where p is the characterization of a data region and f(X) is a linear regression model on a variable of interest y. The novelty of the method lies in the combination of an enumerative approach to explore the space of regions and efficient heuristics that guide the search. Such a strategy provides more flexibility when selecting a small set of jointly accurate and human-readable hybrid rules that explain the entire dataset. As our experiments shows, HiPaR mines fewer rules than existing pattern-based regression methods while still attaining state-of-the-art prediction performance.