I'm doing research comparing boosted decision trees to neural networks for various types of predictive analyses. A boosted decision tree is an ensemble tree created as a series of small trees that form an additive model. I'm using the TreeBoost method of boosting to generate the decision tree series. TreeBoost uses stochastic gradient boosting to increase the predictive accuracy of decision tree models (see http://www.dtreg.com/treeboost.htm).
The available publications comparing boosted trees to neural networks are pretty limited, but the comparisons show boosted trees matching, and in some cases exceeding, the accuracy of neural networks. If you have data that you have successfully (or unsuccessfully) modeled using neural networks, I would like to talk to you. I will be happy to build a boosted decision tree for your data and send you the results so that we can compare the decision tree model to the neural network model. Please e-mail me at phil.sherrod 'at' sandh.com -- Phil Sherrod (phil.sherrod 'at' sandh.com) http://www.dtreg.com (decision tree modeling) http://www.nlreg.com (nonlinear regression) . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
