FlexMF BPR on ML10M

This page analyzes the hyperparameter tuning results for the FlexMF scorer in implicit-feedback mode with pairwise loss (Bayesian Personalized Ranking).

Parameter Search Space

Parameter	Type	Distribution	Values	Selected
embedding_size	Integer	LogUniform	4 ≤ \(x\) ≤ 512	125
regularization	Float	LogUniform	0.0001 ≤ \(x\) ≤ 10	0.187
learning_rate	Float	LogUniform	0.001 ≤ \(x\) ≤ 0.1	0.00238
reg_method	Categorical	Uniform	L2, AdamW	AdamW
negative_count	Integer	Uniform	1 ≤ \(x\) ≤ 5	4
item_bias	Categorical	Uniform	True, False	True

Final Result

Searching selected the following configuration:

{
    'embedding_size': 125,
    'regularization': 0.18690470640031773,
    'learning_rate': 0.002384935477030483,
    'reg_method': 'AdamW',
    'negative_count': 4,
    'item_bias': True,
    'epochs': 12
}

With these metrics:

{
    'RBP': 0.2523367609430753,
    'NDCG': 0.48489473623806856,
    'RecipRank': 0.44162699373801495,
    'TrainTask': '99fe91b6-061b-45a1-a530-075dfa74a2dd',
    'TrainTime': None,
    'TrainCPU': None,
    'max_epochs': 50,
    'done': True,
    'training_iteration': 12,
    'trial_id': '6fcf9_00005',
    'date': '2025-04-04_22-34-50',
    'timestamp': 1743820490,
    'time_this_iter_s': 7.355039834976196,
    'time_total_s': 91.3506453037262,
    'pid': 502774,
    'hostname': 'CCI-ws21',
    'node_ip': '10.248.127.152',
    'config': {
        'embedding_size': 125,
        'regularization': 0.18690470640031773,
        'learning_rate': 0.002384935477030483,
        'reg_method': 'AdamW',
        'negative_count': 4,
        'item_bias': True,
        'epochs': 12
    },
    'time_since_restore': 15.703184127807617,
    'iterations_since_restore': 2
}