Title | ||
---|---|---|
On boosting kernel density methods for multivariate data: density estimation and classification |
Abstract | ||
---|---|---|
Abstract Statistical learning is emerging as a promising fleld where a number,of algorithms from machine learning are interpreted as statistical methods and vice{versa. Due to good practical performance, boosting is one of the most studied machine learning techniques. We propose algorithms for multivariate density estimation and classiflcation. They are generated by using the traditional kernel techniques as weak learners in boosting algorithms. Our algorithms take the form of multistep estimators, whose flrst step is a standard kernel method. Some strategies for bandwidth selection are also discussed with regard both to the standard kernel density classiflcation problem, and to our ‘boosted’ kernel methods. Extensive experiments, using real and simulated data, show an encouraging practical relevance of the flndings. Standard kernel methods are often outperformed by the flrst boosting iterations and in correspondence of several bandwidth values. In addition, the practical efiectiveness of our classiflcation algorithm is conflrmed by a comparative study on two real datasets, the competitors being trees including AdaBoosting with trees. Key words Bandwidth Selection, Bias Reduction, Learning, Leave{One{Out Estimates, Simulation, |
Year | DOI | Venue |
---|---|---|
2005 | 10.1007/s10260-005-0110-1 | Statistical Methods and Applications |
Keywords | Field | DocType |
leave-one-out esti- mates,bandwidth selection,smoothing,learning,simulation,bias reduction,machine learning,density estimation,kernel density,comparative study,kernel method,multivariate data | Econometrics,Semi-supervised learning,Computer science,Unsupervised learning,Artificial intelligence,Ensemble learning,Kernel density estimation,Online machine learning,Multivariate kernel density estimation,Pattern recognition,Boosting (machine learning),Statistics,Machine learning,Gradient boosting | Journal |
Volume | Issue | Citations |
14 | 2 | 8 |
PageRank | References | Authors |
0.58 | 5 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
M. Di Marzio | 1 | 17 | 1.24 |
Charles C. Taylor | 2 | 18 | 4.29 |