VipBoost: a More Accurate Boosting Algorithm
- Xiaoyuan Su
- Taghi Khoshgoftaar, Computer Science and Engineering, Florida Atlantic University, Boca Raton FL 33431, USA
- Russ Greiner, Dept of Computing Science; PI of AICML
Boosting is a well-known method for improving the accuracy of many learning algorithms. In this paper, we propose a novel boosting algorithm, VipBoost (voting on boosting classifications from imputed learning sets), which first generates multiple incomplete datasets from the original dataset by randomly removing a small percentage of observed attribute values, then uses an imputer to fill in the missing values. It then applies AdaBoost (using some base learner) to produce classifiers trained on each of the imputed learning sets, to produce multiple classifiers. The subsequent prediction on a new test case is the most frequent classification from these classifiers. Our empirical results show that VipBoost produces very effective classifiers that significantly improve accuracy for unstable base learners and some stable learners, especially when the initial dataset is incomplete.
Citation
X. Su, T. Khoshgoftaar, R. Greiner. "VipBoost: a More Accurate Boosting Algorithm". Florida AI Research Symposium, pp 356-360, May 2009.Keywords: | machine learning, imputation, boosting |
Category: | In Conference |
BibTeX
@incollection{Su+al:FLAIRS09, author = {Xiaoyuan Su and Taghi Khoshgoftaar and Russ Greiner}, title = {VipBoost: a More Accurate Boosting Algorithm}, Pages = {356-360}, booktitle = {Florida AI Research Symposium}, year = 2009, }Last Updated: July 24, 2009
Submitted by Xiaoyuan Su