CONTROL OF ENERGY EFFICIENCY IN INDUSTRY AND HOUSING AND COMMUNAL SERVICES
Sign | Register
UDC 004.855.5
Methods of decrease in interdependence of errors of decisions trees when training the decisions woods
I.L. Kaftannikov, SUSU
A.V. Parasich, SUSU
Abstract
The problem of training of the decisions woods is considered. The analysis of the existing methods of training of the decisions woods is provided, their advantages and shortcomings are described. Possible methods of decrease in interdependence of errors of trees are offered.
Keywords
decisions trees, decisions woods, machine training, classification
References
1. Vorontsov K.V. [Combinatory Theory of Retraining: Results, Appendices and Open Problems]. Mathematical methods of recognition of images, 2011, pp. 40–43. (in Russ.)
2. Breiman L. Random Forests. Machine Learning, 2001, Vol. 45(1), pp. 5–32.
3. Breiman L. Bagging Predictors. Machine Learning, 1996, Vol. 24, No. 2, pp. 123–140.
4. Y. Freund, R. E. Schapire Experiments with a New Boosting Algorithm. International Conference on Machine Learning, 1996, pp. 148–156.
5. Matsenov A.A. [Committee Busting: Minimization of Number of Basic Algorithms at Simple Vote]. Mathematical methods of recognition of images, 2013, pp. 180–183. (in Russ.)
6. L. Mason, P. Bartlett, J. Baxter Direct Optimization of Margins Improves Generalization in Combined Classifiers. Proc. of the 1998 conf. on Advances in Neural Information Processing Systems II, 1999, pp. 288–294.