MLP-ESE-1Multilayer Perceptron Ensembles trained with Early Stopping.
This learning method trains ensembles of multilayer perceptrons using early stopping. The minimisation algorithm is based on conjugate gradients. The method was implemented by Carl Edward Rasmussen. Both hypertext (generated by LaTeX2HTML) and postscript descriptions of the algorithm are available. A detailed description of the minimization procedure is available in postscript. As well, the source code for the implementation (including the documentation) is available as a compressed tar file.
Results are available from running mlp-ese-1 on most of the Delve datasets. You can either retrieve all of the results for mlp-ese-1 as one large file, or you can retrieve the results for each dataset individually.
You can also get statistics that summarize the results by running mstats, but be warned, the server is very slow. Finally, you may want to jump to the dataset information pages to find out more about them, or what other methods have been run on them.
Results are available for the following datasets:
The kin family of datasets
The pumadyn family of datasets