Open Access

Table 2.

Flare prediction capabilities obtained from our baseline model LR for five different data sets.

Results from our baseline model: Logistic regression

loop24span12 loop24span24 loop24span0 loop12span12 loop48span12
Considering Accuracy 0.95 ± 0.023 0.94 ± 0.007 0.94 ± 0.008 0.96 ± 0.006 0.92 ± 0.010
all Precision (Positive) 0.84 ± 0.023 0.84 ± 0.019 0.79 ± 0.026 0.87 ± 0.021 0.77 ± 0.025
magnetic Precision (Negative) 0.98 ± 0.005 0.98 ± 0.005 0.98 ± 0.005 0.97 ± 0.003 0.97 ± 0.006
parameters Recall (Positive) 0.93 ± 0.018 0.92 ± 0.016 0.93 ± 0.019 0.96 ± 0.014 0.89 ± 0.021
Recall (Negative) 0.95 ± 0.008 0.96 ± 0.006 0.94 ± 0.010 0.97 ± 0.004 0.92 ± 0.011
F1-score (Positive) 0.88 ± 0.014 0.88 ± 0.013 0.85 ± 0.017 0.91 ± 0.014 0.83 ± 0.018
F1-score (Negative) 0.97 ± 0.004 0.97 ± 0.004 0.96 ± 0.005 0.97 ± 0.004 0.94 ± 0.007
HSS1 0.75 ± 0.031 0.74 ± 0.033 0.69 ± 0.041 0.81 ± 0.025 0.64 ± 0.039
HSS2 0.85 ± 0.018 0.85 ± 0.019 0.82 ± 0.022 0.89 ± 0.015 0.77 ± 0.023
GS 0.74 ± 0.026 0.73 ± 0.028 0.68 ± 0.030 0.80 ± 0.024 0.63 ± 0.030
TSS 0.87 ± 0.016 0.87 ± 0.016 0.86 ± 0.018 0.92 ± 0.014 0.81 ± 0.023

Considering Accuracy 0.94 ± 0.006 0.93 ± 0.007 0.94 ± 0.006 0.96 ± 0.005 0.91 ± 0.009
only the top Precision (Positive) 0.83 ± 0.023 0.81 ± 0.022 0.83 ± 0.022 0.85 ± 0.020 0.78 ± 0.025
five magnetic Precision (Negative) 0.97 ± 0.006 0.97 ± 0.006 0.96 ± 0.005 0.98 ± 0.005 0.95 ± 0.008
parameters Recall (Positive) 0.88 ± 0.023 0.87 ± 0.024 0.86 ± 0.023 0.91 ± 0.021 0.84 ± 0.027
Recall (Negative) 0.95 ± 0.008 0.95 ± 0.007 0.96 ± 0.004 0.96 ± 0.006 0.93 ± 0.011
F1-score (Positive) 0.85 ± 0.015 0.84 ± 0.017 0.84 ± 0.014 0.88 ± 0.014 0.81 ± 0.018
F1-score (Negative) 0.96 ± 0.004 0.96 ± 0.005 0.96 ± 0.004 0.97 ± 0.003 0.94 ± 0.006
HSS1 0.69 ± 0.035 0.68 ± 0.036 0.68 ± 0.034 0.75 ± 0.029 0.61 ± 0.041
HSS2 0.81 ± 0.021 0.81 ± 0.022 0.80 ± 0.019 0.85 ± 0.017 0.75 ± 0.024
GS 0.68 ± 0.029 0.67 ± 0.031 0.67 ± 0.028 0.74 ± 0.025 0.60 ± 0.031
TSS 0.83 ± 0.025 0.82 ± 0.023 0.82 ± 0.021 0.87 ± 0.017 0.76 ± 0.027

Notes. The first three data sets, loop24span12, loop24span24, and loop24span0, correspond to the data sets with the same forecasting window of 24 h, but a different span time of 12 h, 48 h, and zero hours, respectively. The last two data sets, loop12span12 and loop48span12, correspond to the data sets with the same span time of 12 h, but have a different forecasting window of 12 h and 48 h, respectively.

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.