Table 6. Open in new tab AUC and F 1... | Oxford Academic

Table 6.

Open in new tab

AUC and F₁ scores given as mean and standard deviation across five runs, trained on the original and augmented training sets with and without redshift, for both with and without transfer learning.

	Model	AUC	\|$\pmb {\mathrm{F}_{1}}$\|
No transfer learning	Original, no redshift	0.891 ± 0.005	0.618 ± 0.005
	Original, with redshift	0.894 ± 0.004	0.59 ± 0.02
	Augmented, no redshift	0.923 ± 0.001	0.677 ± 0.003
	Augmented, with redshift	0.923 ± 0.003	0.651 ± 0.006
With transfer learning	Original, no redshift	0.901 ± 0.002	0.615 ± 0.004
	Original, with redshift	0.915 ± 0.001	0.607 ± 0.009
	Augmented, no redshift	0.924 ± 0.002	0.681 ± 0.001
	Augmented, with redshift	0.946 ± 0.001	0.660 ± 0.006

	Model	AUC	\|$\pmb {\mathrm{F}_{1}}$\|
No transfer learning	Original, no redshift	0.891 ± 0.005	0.618 ± 0.005
	Original, with redshift	0.894 ± 0.004	0.59 ± 0.02
	Augmented, no redshift	0.923 ± 0.001	0.677 ± 0.003
	Augmented, with redshift	0.923 ± 0.003	0.651 ± 0.006
With transfer learning	Original, no redshift	0.901 ± 0.002	0.615 ± 0.004
	Original, with redshift	0.915 ± 0.001	0.607 ± 0.009
	Augmented, no redshift	0.924 ± 0.002	0.681 ± 0.001
	Augmented, with redshift	0.946 ± 0.001	0.660 ± 0.006

Table 6.

Open in new tab

AUC and F₁ scores given as mean and standard deviation across five runs, trained on the original and augmented training sets with and without redshift, for both with and without transfer learning.

	Model	AUC	\|$\pmb {\mathrm{F}_{1}}$\|
No transfer learning	Original, no redshift	0.891 ± 0.005	0.618 ± 0.005
	Original, with redshift	0.894 ± 0.004	0.59 ± 0.02
	Augmented, no redshift	0.923 ± 0.001	0.677 ± 0.003
	Augmented, with redshift	0.923 ± 0.003	0.651 ± 0.006
With transfer learning	Original, no redshift	0.901 ± 0.002	0.615 ± 0.004
	Original, with redshift	0.915 ± 0.001	0.607 ± 0.009
	Augmented, no redshift	0.924 ± 0.002	0.681 ± 0.001
	Augmented, with redshift	0.946 ± 0.001	0.660 ± 0.006

	Model	AUC	\|$\pmb {\mathrm{F}_{1}}$\|
No transfer learning	Original, no redshift	0.891 ± 0.005	0.618 ± 0.005
	Original, with redshift	0.894 ± 0.004	0.59 ± 0.02
	Augmented, no redshift	0.923 ± 0.001	0.677 ± 0.003
	Augmented, with redshift	0.923 ± 0.003	0.651 ± 0.006
With transfer learning	Original, no redshift	0.901 ± 0.002	0.615 ± 0.004
	Original, with redshift	0.915 ± 0.001	0.607 ± 0.009
	Augmented, no redshift	0.924 ± 0.002	0.681 ± 0.001
	Augmented, with redshift	0.946 ± 0.001	0.660 ± 0.006