Table 1 Cross validation results on overcoming the bias of cluster size difference. By using distribution models as reference distribution gap statistics can give correct result even under 10-fold difference.
Difference between sample number of ℙ1 and ℙ2 | Average cluster number estimation accuracy % (Uniform reference distribution) | Average cluster number estimation accuracy % (GMM as reference distribution for ℙ1, uniform reference for ℙ2) |
|---|---|---|
Equal | 100 | 100 |
2-fold | 88.5 | 98.1 |
3-fold | 81.8 | 93.3 |
5-fold | 69.2 | 91.0 |
7-fold | <20 | 89.5 |
9-fold | <20 | 88.9 |
10-fold | <15 | 87.4 |