ACCURACY ANALYSIS OF THE GRADIENT BOOSTING METHOD WITH RANDOM ROTATIONS
https://doi.org/10.21686/2500-3925-2016-4-22-26
Abstract
Gradient boosting method with random rotations is considered, where before training each base learner random rotation is applied to the feature space. The accuracy metric of the given method is estimated for a broad range of generated problems of binary classification. Obtained results are evaluated and recommendations given for application of this method.
About the Author
Victor V. Kitov
Moscow State University;
National Research University “Higher School of Economics”;
Plekhanov Russian University of Economics
Russian Federation
Russian Federation
PhD in Mathematics, mathematician;
docent;
docent
References
1. Hastie T., Tibshirani R., Friedman J. The Elements of Statistical Learning. 2-e izd. – Stanford, USA: Springer, 2009.
2. Abbott D. Why ensembles win data mining competitions. // Predictive Analytics Centre of Excellence Tech Talks, University of California, San Diego. http://pace.sdsc.edu/sites/pace/files/PACE_Abbott_WhyModelEnsemblesWinDataMiningCompetitions_20121114.pdf / 2012.
3. Kitov V.V. Practical aspects of machine learning. // Open systems. SUBD. №1 / 2016. p. 14–17.
4. Chen T., Guestrin C. XGBoost: A Scalable Tree Boosting System. https://arxiv.org/abs/1603.02754. DOI: 10.1145/2939672.2939785.
5. Blaser R., Fryzlewicz P. Random Rotation Ensembles // Journal of Machine Learning Research №17 / 2016. p. 1–26.
6. Ozols M. How to generate a random unitary matrix. http://home.lu.lv/~sd20008/papers/essays/Random%20unitary%20[paper].pdf /2009.
Review
For citations:
Kitov V.V. ACCURACY ANALYSIS OF THE GRADIENT BOOSTING METHOD WITH RANDOM ROTATIONS. Statistics and Economics. 2016;(4):22-26. (In Russ.) https://doi.org/10.21686/2500-3925-2016-4-22-26
Views:
808
.png)
.png)
Email this article 