Improved Uniformity Enforcement in Stochastic Discrimination
There are a variety of methods for inducing predictive systems from observed data. Many of these methods fall into the field of study of machine learning. Some of the most effective algorithms in this domain succeed by combining a number of distinct predictive elements to form what can be described as a type of committee. Well known examples of such algorithms are AdaBoost, bagging and random forests. Stochastic discrimination is a committee-forming algorithm that attempts to combine a large number of relatively simple predictive elements in an effort to achieve a high degree of accuracy. A key element of the success of this technique is that its coverage of the observed feature space should be uniform in nature. We introduce a new uniformity enforcement method, which on benchmark datasets, leads to greater predictive efficiency than the currently published method.
KeywordsFeature Space Random Forest Test Error Normalise Standard Deviation Coverage Array
Unable to display preview. Download preview PDF.
- 1.Asuncion, A., Newman, D.J.: UCI machine learning repository (2007)Google Scholar
- 3.Breiman, L.: Bias, variance, and arcing classifiers (1996)Google Scholar
- 5.Freund, Y., Schapire, R.E.: Experiments with a new boosting algorithm. In: International Conference on Machine Learning, pp. 148–156 (1996)Google Scholar
- 6.Garner, S.R.: Weka: The waikato environment for knowledge analysis. In: Proc. of the New Zealand Computer Science Research Students Conference, pp. 57–64 (1995)Google Scholar
- 8.Kleinberg, E.M.: Stochastic discrimination. Annals of Mathematics and Artificial Intelligence 1 (1990)Google Scholar
- 10.Kleinberg, E.M., Ho, T.K.: Pattern recognition by stochastic modeling. In: Proceedings of the Third International Workshop on Frontiers in Handwriting Recognition, pp. 175–183. Partners Press (1993)Google Scholar
- 12.Prior, M., Windeatt, T.: Parameter tuning using the out-of-bootstrap generalisation error estimate for stochastic discrimination and random forests. In: International Conference on Pattern Recognition, pp. 498–501 (2006)Google Scholar