Table 1. za,g and t parameter values for which the highest R2 values were obtained when correlating the training subsets of the log(FN_mig_cori = ra,g,o,d,p = 1) and the log(acsa,s,o,d).
We performed four rounds of cross-validation by randomly selecting in each round 51 countries of birth as the training subset and the remaining 25 countries of birth as the testing subset.
g | a | Training subsets, 51 countries of birth in the US | Median | Training with Eurostat [8] | ||||
---|---|---|---|---|---|---|---|---|
1a | 2a | 3a | 4a | |||||
za,g | female | 15–24 | 0.8 | 1 | 0.9 | 0.8 | 0.85 | 0.85 |
25–39 | 1 | 0.8 | 0.7 | 0.7 | 0.75 | 0.5 | ||
40–54 | 0.1 | 0.3 | 0.4 | 0.3 | 0.3 | 0.3 | ||
55–64 | 0 | 0.3 | 0.3 | 0.2 | 0.25 | 0.3 | ||
male | 15–24 | 0.7 | 1 | 0.9 | 0.9 | 0.9 | 0.95 | |
25–39 | 1 | 1 | 0.8 | 1 | 1 | 0.8 | ||
40–54 | 0.4 | 0.6 | 0.6 | 0.6 | 0.6 | 0.8 | ||
55–64 | 0.5 | 0.9 | 0.6 | 0.6 | 0.6 | 0.85 | ||
t | 0.6 | 0.2 | 0.2 | 0.3 | 0.25 | 0.25 | ||
Training Max R2 | 0.92 | 0.90 | 0.90 | 0.91 | 0.91 | 0.83 | ||
Testing subsets, 25 countries of origin in the US | ||||||||
1b | 2b | 3b | 4b | |||||
Testing R2 | 0.87 | 0.90 | 0.90 | 0.86 |