Table 4.
Number of clusters comparisons with known CV, p0 = 0.2 and p1 = 0.3
| ρ | n | CV | Li and Redden[1] | MD-corrected estimator[2] | FG-corrected estimator with d=0.1[2] | FG-corrected estimator with d=0.75[2] |
|---|---|---|---|---|---|---|
| 0.01 | 140 | 0.1–0.4 | 16 | 20 | 18 | 20 |
| 0.5 | 16 | 28 | 22 | 26 | ||
| 0.6 | 18 | |||||
| 200 | 0.1–0.4 | 14 | 18 | 16 | 18 | |
| 0.5–0.6 | 16 | 26 | 20 | 24 | ||
| 0.03 | 140 | 0.1–0.3 | 28 | 30 | 30 | 30 |
| 0.4 | 30 | |||||
| 0.5 | 32 | 38 | 36 | 38 | ||
| 0.6 | 34 | |||||
| 200 | 0.1–0.2 | 26 | 30 | 30 | 30 | |
| 0.3–0.4 | 28 | |||||
| 0.5 | 30 | 38 | 36 | 38 | ||
| 0.6 | 32 |
Assume unequal cluster sizes;
Adjustment from the proposed algorithm due to unequal cluster sizes for finite clusters;