Skip to main content
. 2009 Jun 19;25(21):2882–2889. doi: 10.1093/bioinformatics/btp378

Fig. 7.

Fig. 7.

The performance of ISOLATE, LDA and another Kullback-Leibler (KL) divergence-based measure on the clinical dataset of 93 tumor samples. Each sample is predicted independently of all other samples in the dataset. The number of samples in each class is shown beside the class name, and classes are in decreasing order of size, from left to right, with the overall performance shown in the leftmost column. The black line shows random performance.