kiez.evaluate.eval_metrics

Calculate evaluation metrics such as hits@k.

Functions

hits(nn_ind, gold[, k])

Show hits@k.

kiez.evaluate.eval_metrics.hits(nn_ind: Union[ndarray, Dict[Any, List]], gold: Dict[Any, Any], k=None) → Dict[int, float]

Show hits@k.

Parameters:

• nn_ind (array or dict) – Contains the indices of the nearest neighbors for source entities.

• gold (dict) – Map of source indices to the respective target indices

• k (array) – Which ks should be evaluated

Returns:

hits – k: relative hits@k values

Return type:

dict

Examples

>>> from kiez.evaluate import hits
>>> import numpy as np
>>> nn_ind = np.array([[1, 2, 3], [2, 3, 4], [3, 4, 5], [4, 5, 6]])
>>> gold = {0: 2, 1: 4, 2: 3, 3: 4}
>>> hits(nn_ind, gold)
{1: 0.5, 5: 1.0, 10: 1.0}