ctf4science.eval_module.short_time_forecast

ctf4science.eval_module.short_time_forecast(truth: ndarray, prediction: ndarray, k: int) → float

Compute the short-time forecast score (relative L2 over first k steps, as percentage).

Uses the first k time steps of truth and prediction; score is 100 * (1 - relative_L2_error) so that 100 is perfect.

Parameters:

truth(T, F) ndarray

Ground truth, shape (time steps, features).

prediction(T, F) ndarray

Predicted data, same shape as truth.

kint

Number of initial time steps to use.

Returns:

float

Short-time forecast score as a percentage (100 = perfect).