LAFTR continuous label demographic parity test:

gamma = 0
MSE = 110.2866850
DP = 5.6777613
D_EO = 0.857
D_err = 13.84

gamma = 0.1
MSE = 110.5425
DP = 5.0826
D_EO = 1.4522
D_err = 12.379

gamma = 1
MSE = 114.804
DP = 2.267
D_EO = 4.268
D_err = 4.4599

gamma = 4
MSE = 119.2566
DP = 1.2039
D_EO = 5.3309
D_err = 1.8630

gamma = 16
MSE = 128.38
DP = 0.0511
D_EO = 6.5859
D_Err = 6.9569

gamma = 64
MSE = 138.76
DP = 1.166
D_EO = 5.36877
D_err = 5.714806


STORKEY Hours worked(continuous) label demographic parity test:

gamma = 0
MSE = 110.2866850
DP = 5.6777613
D_EO = 0.857
D_err = 13.84
gamma = 0.1
MSE = 110.3463832
DP = 5.6431743
D_EO = 0.8917
D_err = 13.4768

gamma = 1
MSE = 110.7191588
DP = 5.3605390
D_EO = 1.1743
D_err = 13.57

gamma = 4
MSE = 109.8540942
DP = 5.2137214
D_EO = 1.3211
D_err = 12.911

gamma = 16
MSE = 110.7215910
DP = 4.4828112
D_EO = 2.052
D_err = 12.86
gamma = 64
MSE = 113.0438344
DP = 2.9873861
D_EO = 3.55
D_err = 7.728



LAFTR continuous sens variable demographic parity test:
Validation done by choosing lowest error + bin variance

gamma = 0
Acc = 0.8395
inv DP = 7.701
log-reg = 0.0408
bin var = 0.1214

gamma = 0.0001
Acc = 0.8431
inv DP = 7.73
log-reg = 0.0411
bin-var = 0.121

gamma = 0.0005
Acc = 0.8400
inv DP = 7.358
log-reg = 0.0387
bin-var = 0.1161

gamma = 0.001
Acc = 0.8449
inv DP = 7.674
log-reg = 0.04053
bin-var = 0.1220

gamma = 0.01
Acc = 0.8172
inv DP = 6.723
log-reg = 0.0341
bin-var = 0.0789

gamma = 0.1
Acc = 0.7799
inv DP = 2.905
log-reg = 0.012689
bin-var = 0.0324

STORKEY continuous sens variable demographic parity test:


gamma = 0
Acc = 0.8395
inv DP = 7.701
log-reg = 0.0408
bin var = 0.1214

gamma = 0.0001
Acc = 0.8317
inv DP = 8.585
log-reg = 0.0455
bin-var = 0.1026

gamma = 0.0005
Acc = 0.8181
inv DP = 3.786
log-reg = 0.01899
bin-var = 0.0547

gamma = 0.001
Acc = 0.7560
inv DP = 5.704
log-reg = 0.0941
bin-var = 0.001

gamma = 0.01
Acc = 0.758
inv DP = nan** (Y_hat = 0 for all predictions, so maybe use inv DP = 0 for this?)
log-reg = 0
bin-var = 0