Skip to main content

Table 3 Performance of SVM and classification trees.

From: Reference-free inference of tumor phylogenies from single-cell sequencing data

Classifier

Class Distribution

Min

Median

Mean

Max

Standard

deviation

SVM

Uneven, all samples

Even, subsampled

0.003

0.003

0.005

0.02

0.006

0.02

0.01

0.03

0.002

0.008

CART

Uneven, all samples

Even, subsampled

0.004

0

0.005

0.016

0.008

0.02

0.016

0.04

0.002

0.013

  1. Values shown are classification errors computed as the average ratio of incorrectly classified test samples to the total number of test samples per replicate. We report results both including the entire dataset and results from 10 replicates of subsampling to ensure an even class distribution.