sliding_window¶

sliding_window(x_data: ndarray, labels: ndarray, scorer: callable, window_size: int = 20, axis: int = -1, n_jobs: int = -3, **kwargs) → ndarray[source][source]¶

Compute a function over a sliding window.

Parameters:

x_data (np.ndarray, shape (..., trials, time)) – The data to compute the function over
labels (np.ndarray, shape (trials,)) – The labels for each trial
scorer (callable) – The function to compute over the sliding window. Must take two arguments, the data and the labels.
window_size (int) – The size of the sliding window
axis (int) – The axis to compute the sliding window over
n_jobs (int) – The number of jobs to run in parallel

Returns:

The output of the function, shape (…, time - window_size + 1)

Return type:

np.ndarray

Examples

>>> def square(x, labels):
...     return np.mean(x ** 2, where=labels == 1)
>>> x_data = np.arange(40).reshape(4, 10)
>>> labels = np.array([0, 1, 1])
>>> sliding_window(x_data, labels, square, window_size=3)
array([397.5, 431.5, 467.5, 505.5, 545.5, 587.5, 631.5])