TABLE 2.
Medians of the differences in performance between balance selection methodsa
| Method | Median of difference in performance of indicated method |
|||||
|---|---|---|---|---|---|---|
| selbal | PBA-distal | ABA-distal | RBA-distal | DBA-distal | DBA | |
| selbal | −0.125 to −0.091 | −0.012 to 0.016 | −0.0066 to 0.0203 | 0.016 to 0.042 | 0.019 to 0.045 | |
| PBA-distal | 0.091 to 0.125 | 0.087 to 0.122 | 0.095 to 0.131 | 0.12 to 0.16 | 0.13 to 0.16 | |
| ABA-distal | −0.016 to 0.012 | −0.122 to −0.087 | −0.0082 to 0.0182 | 0.014 to 0.040 | 0.017 to 0.044 | |
| RBA-distal | −0.0203 to 0.0066 | −0.131 to −0.095 | −0.0182 to 0.0082 | 0.0082 to 0.0345 | 0.012 to 0.038 | |
| DBA-distal | −0.042 to −0.016 | −0.16 to −0.12 | −0.040 to −0.014 | −0.0345 to −0.0082 | −0.006 to 0.014 | |
| DBA | −0.045 to −0.019 | −0.16 to −0.13 | −0.044 to −0.017 | −0.038 to −0.012 | −0.014 to 0.006 | |
Confidence intervals computed using pairwise Wilcoxon rank sum tests applied to 50 resamplings of 13 data sets. The appendix “-distal” indicates that only the 2-part and 3-part balances were used as features. This table corresponds to Fig. 3.