Table 5 Split of datasets of the predictor.

From: Disruption prediction for future tokamaks using parameter-based transfer learning

 

Training

Validation

Test

Time threshold (ms)

J-TEXT

494 (189)

140 (70)

220 (110)

62

E AST (all data)

1896 (355)

120 (60)

360 (180)

172

EAST (transfer)

20 (10)

/

360 (180)

172

  1. Values in parentheses give the number of disruptive discharges. The time threshold is used to define disruptive versus non-disruptive labels.
Back to article page