Skip to main content
Figure 4 | BMC Bioinformatics

Figure 4

From: Unsupervised assessment of microarray data quality using a Gaussian mixture model

Figure 4

Classifier Performance. Unsupervised versus supervised classifier using labeled data sets of various sizes. When the full labeled training dataset (~540 labeled instances per fold) is available, the performance of the unsupervised classification method (EM+Naïve Bayes) and the supervised classification method (MLE+Naïve Bayes) are equivalent on the test dataset. When the amount of labeled data is limited, but unlabeled data is abundant, the unsupervised method outperforms the supervised method.

Back to article page