Research Article

Effects of Pooling Samples on the Performance of Classification Algorithms: A Comparative Study

Figure 3

The performance of classifiers for various numbers of markers based on non-pooled and pooled data. Misclassification rates of classifiers are shown when the number of markers is increasing from 1 to 10. (a) shows the performance of classifiers on dataset of individual samples (90 controls and 90 cases). (b)–(d) show the performance of classifiers on pooled datasets when pool sizes are 2, 3, and 5, respectively.
278352.fig.003a
(a)
278352.fig.003b
(b)
278352.fig.003c
(c)
278352.fig.003d
(d)