Using bundle embeddings to predict daily cortisol levels in human subjects

Table 2 SMAP results for the full embedding

ID	dim	n	nrmse_min	∆nrmse	θ
1	1	107	0.689	0.001	0.2
2	1	114	1.015	0.000	0.0
3	1	112	0.323	0.021^b	1.0
4	1	115	0.762	0.004^b	0.2
5	1	119	0.385	0.045^b	1.0
6	1	112	0.538	0.014^b	0.6
7	3	101	0.426	0.080^b	0.8
8	1	111	0.560	0.002^a	0.2
9	1	116	0.584	0.007^b	0.4
10	1	110	0.410	0.006^b	0.6

Note: ID = subject ID, dim = dimension of the embedding, n = number of vectors in the embedding, nrmse_min = minimum of the NRMSE, Δnrmse = difference between the minimal NRMSE (θ > = 0) and the NRMSE at θ = 0 (the global linear case), θ = value of θ where the NRMSE was minimal
^aSignificant at the 0.05 level
^bSignificant at the 0.01 level

ISSN: 1471-2288