Winning the KDD Cup Orange Challenge with ensemble selection
作者： Claudia Perlich , Vikas Sindhwani , Alexandru Niculescu-Mizil , Prem Melville , Grzegorz Swirszcz
DOI:
关键词:
摘要: … a classifier library is generated, Ensemble Selection builds an ensemble by selecting from the library … So, for appetency, we decided to use the ensemble model generated right after the …
uchicago.edu 
niculescu-mizil.org 参考文章(23)
Chad M. Cumby, Dan Roth, Jeff L. Rosen, Andrew J. Carlson, SNoW User Guide University of Illinois at Urbana-Champaign. ,(1999)
Tomaso Poggio, Ryan Michael Rifkin, Everything old is new again: a fresh look at historical approaches in machine learning Everything old is new again: a fresh look at historical approaches in machine learning. pp. 1- 1 ,(2002)
Michael W. Berry, SVDPACK: A Fortran-77 Software Library for the Sparse Singular Value Decomposition University of Tennessee. ,(1992)
Mark A. Hall, Ian H. Witten, Eibe Frank, Data Mining: Practical Machine Learning Tools and Techniques ,(1999)
Dan Roth, Learning to resolve natural language ambiguities: a unified approach national conference on artificial intelligence. pp. 806- 813 ,(1998)
George H. John, Pat Langley, Estimating continuous distributions in Bayesian classifiers uncertainty in artificial intelligence. pp. 338- 345 ,(1995)
Rich Caruana, Alexandru Niculescu-Mizil, Geoff Crew, Alex Ksikes, Ensemble selection from libraries of models Twenty-first international conference on Machine learning - ICML '04. pp. 18- ,(2004) , 10.1145/1015330.1015432
Nicholas Littlestone, Redundant noisy attributes, attribute errors, and linear-threshold learning using winnow conference on learning theory. pp. 147- 156 ,(1991) , 10.5555/114836.114850
Jyrki Kivinen, Manfred K. Warmuth, EXPONENTIATED GRADIENT VERSUS GRADIENT DESCENT FOR LINEAR PREDICTORS Information & Computation. ,vol. 132, pp. 1- 63 ,(1997) , 10.1006/INCO.1996.2612
Thorsten Joachims, A support vector method for multivariate performance measures Proceedings of the 22nd international conference on Machine learning - ICML '05. pp. 377- 384 ,(2005) , 10.1145/1102351.1102399