Tree-based approaches to analytics have been the most powerful and widely used methods in Business Intelligence for several years.  Tree-based approaches are the most used in predictive analytics due to their unmatched ability to:

• Account for non-linear effects
• Quickly and easily elucidate complex substructure such as in demographic or scientific data
• Deliver truly non-biased, statistically significant analysis of highly dimensional, seemingly unrelated data.

Further, results from tree-based analytics are typically easier to use and understand versus competing methods, and they generate results that are both logical and visually interpretable. The tools themselves tend to be easier to use, yet they enable the analysis of more dimensionally complex data – meaning a greater number of variables – thus resulting in more predictive results.

In short, tree-based tools are easier to use, more versatile, and typically more powerful than competing methods.

Optimus RP uses the most advanced form of tree-based modeling, known as FIRMplus, which is based on Formal Inference-based Recursive Modeling, or FIRM.  FIRM was developed by Professor Douglas Hawkins, PhD, Chair of Applied Statistics at the University of Minnesota’s School of Statistics.  As the name suggests, FIRMplus uses the conceptual framework of FIRM, but with major enhancements in both the analysis capability and user interface.

Within FIRM, a recursive partitioning engine attempts to find a single predictor whose different values split a given covariate into subgroups based on true statistical significance.  Recursive partitioning is more flexible than methods like multiple regression, and allows any sort of relationship between a predictor and the response, not just a straight line, and will use different predictors in different parts of the tree, reflecting the fact that one predictor may be relevant in one subgroup, with a different predictor being more relevant in another subgroup.