next up previous
Next: About this document ... Up: Popular Ensemble Methods: An Previous: Bibliography

Appendix

Tables 4 and 5 show the complete results for the first set of experiments used in this paper.

  
Table 4: Neural network test set error rates and standard deviation values for those error rates for (1) a single neural network classifier; (2) a simple neural network ensemble; (3) a Bagging ensemble; (4) an Arcing ensemble; and (5) and Ada-Boosting ensemble. Also shown (results column 3) is the ``best'' result produced from all of the single network results run using all of the training data.
\begin{table}
\vspace{-0.1in}
\begin{tabular*}{\textwidth}{@{\extracolsep{\fill}...
... & 19.1 & 1.0 & 19.7 & 1.0 \\ \hline
\end{tabular*}
\vspace{-0.15in}\end{table}


  
Table 5: Decision tree test set error rates and standard deviation values for those error rates for (1) a single decision tree classifier; (2) a Bagging ensemble; (3) an Arcing ensemble; and (4) and Ada-Boosting ensemble. Also shown (results column 3) is the ``best'' result produced from all of the single tree results run using all of the training data.
\begin{table}
\vspace{-0.1in}
\begin{tabular*}{\textwidth}{@{\extracolsep{\fill}...
... & 19.1 & 1.0 & 19.7 & 1.0 \\ \hline
\end{tabular*}
\vspace{-0.15in}\end{table}



David Opitz
1999-08-24