The American Marketing Association website singles out this poster as a well-deserved award winner: "Random forests - applied to treatment decision making"

The application of Random Forests as an alternative to other data mining techniques has been shown to have many benefits. This paper demonstrates the usefulness of this technique to modeling the variable distributed (physician and patient) treatment decision processes involved in healthcare.

The benefits of using Random Forests are clearly shown through 3D surface charts and performance measures. Although there are some broad similarities, the ability of Random Forests to form clear cut off points for where a treatment is prescribed, and the capacity to model interaction effects are clearly identified.

Return to news index
Return to home page