The DataModeler package was motivated by the recognition that the industry is great at collecting data ... and then performing records retention and discarding that data. In between, we really ought to do something to extract the value from that data; however, extracting insight from multivariate data is intrinsically hard.

Fortunately, the intersection of emerging concepts (such as genetic programming symbolic regression) with improvements in enabling technologies (such as faster and more powerful computers as well as advances in algorithms and analysis infrastructures such as Mathematica) allow us to effectively attack data modeling and analysis problems which would have been intractable in the recent past.

Since the human time investment is generally the most expensive aspect of empirical model building, the DataModeler package in conjunction with Mathematica provides an infrastructure for human-centric data exploration, model development, model exploration and model management and exploitation. Of course, given the difficulty of nonlinear data modeling, we also put an emphasis on efficient algorithms at the core of this modeling infrastructure.

Pages below provide more information about industrial-strength data driven modeling and motivation behind it as well as the Quick Start Tutorials that will help you get a grip of the DataModeler: