Name
Abstract | ||
|---|---|---|
In this paper we propose a novel cost function to train a standard SLP, which proves to perform similarly to a linear SVM, currently considered as one of the best linear discriminants. Then we show how we can use regularized boosting to construct a conventional MLP which does not suffer from overfitting, and needs no adjustment in the number of hidden units. This MLP consistently outperforms carefully constructed regularized and non-regularized MLPs trained using backpropagation, nonlinear SVMs and sometimes even MLP ensembles (constructed with Real Adaboost-like algorithms), with the advantage of a much simpler structure. To define each of these MLP models, just two parameters are needed, the booster regularization parameter and A, which accounts for regularization in the learners. |
Year | Venue | Keywords |
|---|---|---|
2006 | PROCEEDINGS OF THE SECOND IASTED INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE | boosting, cost functions, Parzen models |
Field | DocType | Citations |
Pattern recognition,Computer science,Speech recognition,Artificial intelligence,Boosting (machine learning),Machine learning | Conference | 0 |
PageRank | References | Authors |
0.34 | 1 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Miguel Lázaro-Gredilla | 1 | 383 | 26.46 |
| Jaisiel Madrid-Sánchez | 2 | 0 | 1.01 |
| Aníbal R. Figueiras-Vidal | 3 | 467 | 38.03 |