def basic_usage_large_dataset():
from recommender_metrics import calculate_metrics
from recommender_metrics import generate_random_data
import json
groups, scores, labels = generate_random_data(n_users=50000)
print("Larger data:")
print(" #groups:", len(groups))
print(" #scores:", len(scores))
print(" #labels:", len(labels))
print()
metrics = calculate_metrics(group_ids=groups, scores=scores, labels=labels)
print("Metrics:")
print(json.dumps(metrics, indent=2))
print("\n\n\n")
def basic_usage_group_filttering():
from recommender_metrics import calculate_metrics
from recommender_metrics import generate_random_data
import json
print("Running example3")
groups, scores, labels = generate_random_data()
print("Data:")
print(" #groups:", len(groups))
print(" #scores:", len(scores))
print(" #labels:", len(labels))
print()
metrics = calculate_metrics(group_ids=groups,
scores=scores,
labels=labels,
remove_empty=True)
print("Metrics:")
print(json.dumps(metrics, indent=2))
print("\n\n\n")
def test_empty(self):
target = {
"mAP@1": 0.2631578947368421,
"precision@1": 0.2631578947368421,
"recall@1": 0.07631578947368421,
"mAP@5": 0.5046783625730993,
"precision@5": 0.3140350877192983,
"recall@5": 0.45743525480367586,
"mAP@10": 0.4538314536340852,
"precision@10": 0.2941520467836258,
"recall@10": 0.6738512949039264,
"mAP@20": 0.424754158736318,
"precision@20": 0.33355881010883454,
"recall@20": 1.0,
}
groups, scores, labels = recommender_metrics.generate_random_data()
metrics = recommender_metrics.calculate_metrics(group_ids=groups,
scores=scores,
labels=labels,
remove_empty=True)
self.dict_vals_all_close(target=target,
pred=metrics,
desc=f"Removal of empty group labels")
def basic_usage_custom_metrics_k():
from recommender_metrics import calculate_metrics
from recommender_metrics import generate_random_data
import json
print("Running example4")
groups, scores, labels = generate_random_data()
print("Data:")
print(" #groups:", len(groups))
print(" #scores:", len(scores))
print(" #labels:", len(labels))
print()
metrics = calculate_metrics(
group_ids=groups,
scores=scores,
labels=labels,
k_list=[1, 2, 4, 8, 16],
metrics=["mAP", "precision", "recall", "ndcg", "auroc"],
)
print("Metrics:")
print(json.dumps(metrics, indent=2))
print("\n\n\n")
"recall@1": 0.1,
"mAP@5": 0.8041666666666667,
"precision@5": 0.8,
"recall@5": 0.4,
"mAP@10": 0.8121315192743763,
"precision@10": 0.7,
"recall@10": 0.7,
"mAP@20": 0.7555050505050506,
"precision@20": 0.5,
"recall@20": 1.0,
},
),
dict(
kwargs=dict(
zip(("group_ids", "scores", "labels"),
recommender_metrics.generate_random_data())),
name="Bigger random sample",
targets={
"mAP@1": 0.25,
"precision@1": 0.25,
"recall@1": 0.12250000000000001,
"mAP@5": 0.47944444444444434,
"precision@5": 0.2983333333333334,
"recall@5": 0.4845634920634921,
"mAP@10": 0.4311398809523809,
"precision@10": 0.2794444444444445,
"recall@10": 0.69015873015873,
"mAP@20": 0.4035164507995021,
"precision@20": 0.3168808696033928,
"recall@20": 1.0,
},