NPMR is a data mining algorithm that can automatically accommodate complex interactions to enable prediction with complicated data; NPMR was originally developed for ecological data (McCune 2006). NPMR is well-suited to non-linear data that have more than one factor influencing a response of interest. NPMR can detect unusual shapes in data such as thresholds or other irregular but important shapes.
See a paper comparing NPMR to other common data mining algorithms here (Lintz et al. 2011). The paper develops a conceptual framework for testing data mining algorithms and implements it. It develops 48 simulated data sets for use in algorithm testing and comparison, and these are available from the supplemental material on the journal's website.
Lintz., H., McCune, B., Gray, A., and McCulloh, K. 2011. Quantifying ecological thresholds from response surfaces. Ecological Modelling 222: 427-436.
McCune, B., 2006. Non-parametric habitat models with automatic interactions. Journal of Vegetation Science 17: 819-830.