Logging

utils.logs.log_results(path: str, features: list[int], loss: list[float], accuracy: list[float], periter: dict[slice(<class 'str'>, list[int], None)], eval_times: list[float], transform_time: float) → None

Logs AHFS results from a single run.

Parameters:

• path (str) – Folder path to establish logs in, must not exist. When running AHFS multiple times, establish a superdirectory, e.g. “novelAHFS_runs/{dataset}”. Epoch time is automatically concatenated to the end of the path to avoid overwrite. Folder naming should follow the format {arbitrary text}_{dataset name}_{arbitrary text} to ensure compatibility with other utility functions.

• features (list[int]) – Ordered list of selected features.

• loss (list[float]) – Ordered list of loss per iteration.

• accuracy (list[float]) – Ordered list of accuracy per iteration.

• periter (dict[str: list[int]]) – Dictionary with measure names as keys and ordered list of candidate features suggested by the measure.

• eval_times (list[float]) – Ordered list of evaluation time per iteration.

• transform_time (float) – Transform time.

Returns:

None