Application of neural networks to synchro-Compton blazar emission models

Open Access

Table 3

Evaluation of results.

	Cell type

	ANN	GRU	LSTM
Metric	Mean score^(*)

R²	0.902	0.985	0.989
MSE	1.84	0.23	0.20
MAE	0.89	0.13	0.13
KS	0.256	0.070	0.068
DTW	8.86	3.12	4.37

Ranking percentages (%)^(**)

MSE	0/1.3/98	42/56/1	57/42/0
MAE	0/0/100	79/21/0	21/79/0
KS	0.5/2.9/97	54/44/1.8	46/53/1.6
DTW	0.00/1.2/99	87/12/0.7	13/86/0.4

Notes.^(*) Median metric value for each NN based on the test sample. ^(**)Percentage of times where each NN ranked 1st/2nd/3rd in their between comparison for the test sample.

Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.

Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.

Initial download of the metrics may take a while.