deepmol.metrics package

Submodules

deepmol.metrics.metrics module

class Metric(metric: callable, task_averager: callable | None = None, name: str | None = None, **kwargs)

Bases: object

Class for computing machine learning metrics.

Metrics can be imported from scikit-learn or can be user defined functions.

compute_metric(y_true: ndarray, y_pred: ndarray, n_tasks: int, per_task_metrics: bool = False) → Tuple[Any, List[Any]]

Compute a performance metric for each task.

Parameters:

• y_true (np.ndarray) – An np.ndarray containing true values for each task.

• y_pred (np.ndarray) – An np.ndarray containing predicted values for each task.

• n_tasks (int) – The number of tasks this class is expected to handle.

• per_task_metrics (bool) – If true, return computed metric for each task on multitask dataset.

• kwargs (dict) – Will be passed on to self.metric

Returns:

Tuple with the task averager computed value and a list containing metric values for each task.

Return type:

Tuple[Any, List[Any]]

compute_singletask_metric(y_true: ndarray, y_pred: ndarray, **kwargs) → float

Compute a metric value for a singletask problem.

Parameters:

• y_true (np.ndarray) – True values array.

• y_pred (np.ndarray) – Predictions array.

• kwargs (dict) – Will be passed on to self.metric

Returns:

metric_value – The computed value of the metric.

Return type:

float

deepmol.metrics.metrics_functions module

Script containing imports of metrics and new metric functions.

pearson_score(y, y_pred)

prc_auc_score(y, y_pred)

spearman_score(y, y_pred)

Module contents