%0 Journal Article %1 friedman_stochastic_2002 %A Friedman, Jerome H. %D 2002 %J Computational Statistics & Data Analysis %K Boosting, Learning, Machine sampling %N 4 %P 367--378 %R 10.1016/S0167-9473(01)00065-2 %T Stochastic gradient boosting %U http://www.sciencedirect.com/science/article/pii/S0167947301000652 %V 38 %X Gradient boosting constructs additive regression models by sequentially fitting a simple parameterized function (base learner) to current “pseudo”-residuals by least squares at each iteration. The pseudo-residuals are the gradient of the loss functional being minimized, with respect to the model values at each training data point evaluated at the current step. It is shown that both the approximation accuracy and execution speed of gradient boosting can be substantially improved by incorporating randomization into the procedure. Specifically, at each iteration a subsample of the training data is drawn at random (without replacement) from the full training data set. This randomly selected subsample is then used in place of the full sample to fit the base learner and compute the model update for the current iteration. This randomized approach also increases robustness against overcapacity of the base learner.

@article{friedman_stochastic_2002, abstract = {Gradient boosting constructs additive regression models by sequentially fitting a simple parameterized function (base learner) to current “pseudo”-residuals by least squares at each iteration. The pseudo-residuals are the gradient of the loss functional being minimized, with respect to the model values at each training data point evaluated at the current step. It is shown that both the approximation accuracy and execution speed of gradient boosting can be substantially improved by incorporating randomization into the procedure. Specifically, at each iteration a subsample of the training data is drawn at random (without replacement) from the full training data set. This randomly selected subsample is then used in place of the full sample to fit the base learner and compute the model update for the current iteration. This randomized approach also increases robustness against overcapacity of the base learner.}, added-at = {2017-01-09T13:57:26.000+0100}, author = {Friedman, Jerome H.}, biburl = {https://www.bibsonomy.org/bibtex/29eb90a5284c226df3f5e4b342aa4a3a3/yourwelcome}, doi = {10.1016/S0167-9473(01)00065-2}, interhash = {49094303fc87fbfa18cc649acd2c7e95}, intrahash = {9eb90a5284c226df3f5e4b342aa4a3a3}, issn = {0167-9473}, journal = {Computational Statistics \& Data Analysis}, keywords = {Boosting, Learning, Machine sampling}, month = feb, number = 4, pages = {367--378}, shorttitle = {Nonlinear {Methods} and {Data} {Mining}}, timestamp = {2017-01-09T14:01:11.000+0100}, title = {Stochastic gradient boosting}, url = {http://www.sciencedirect.com/science/article/pii/S0167947301000652}, urldate = {2013-06-04}, volume = 38, year = 2002 }