Loading...
Please wait, while we are loading the content...
Similar Documents
Linear Time Nonparametric Classification and Feature Selection with Polynomial MPMC Cascades for Large Datasets ; CU-CS-977-04
| Content Provider | Semantic Scholar |
|---|---|
| Author | Breitenbach, Markus Bohte, Sander M. Grudic, Gregory Z. |
| Copyright Year | 2004 |
| Abstract | The recently proposed Polynomial MPMC Cascade (PMC) algorithm is a nonparametric classifier for high-dimensional non-linear binary classification with performance competitive with state-of-the-art classifiers like SVMs. Importantly, the algorithm has linear-time complexity with respect to both training-set size and dimensionality of the problem. In this paper, we show how we can exploit this computational efficiency to build classifiers from very large datasets typical in datamining problems. Furthermore, we demonstrate that the PMC algorithm efficiently does feature selection in such very large large problem domains. Experimental results are given on datasets ranging in sizes between 170,000 and 4.8 million examples. We empirically verify the linear time dependence of the algorithm, and explore how the stability of the classifier is influenced by sample size. The techniques discussed in this paper should allow nonlinear binary classifiers to be efficiently learned from tens of millions of training data, with feature selection being a |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://www.cs.colorado.edu/department/publications/reports/docs/CU-CS-977-04.pdf |
| Alternate Webpage(s) | http://markus-breitenbach.com/download/cascade_2.pdf |
| Alternate Webpage(s) | https://scholar.colorado.edu/cgi/viewcontent.cgi?article=1914&context=csci_techreports&httpsredir=1&referer= |
| Alternate Webpage(s) | http://scholar.colorado.edu/cgi/viewcontent.cgi?article=1914&context=csci_techreports |
| Alternate Webpage(s) | http://www.cs.colorado.edu/~grudic/publications/cascade_2.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
