This page analyzes the hyperparameter tuning results for the implicit-feedback ALS matrix factorization model.
Parameter Search Space
/home/mde48/lenskit/lenskit-codex/.venv/lib/python3.12/site-packages/ray/tune/search/sample.py:700: RayDeprecationWarning: The `base` argument is deprecated. Please remove it as it is not actually needed in this method.
| embedding_size |
Integer |
LogUniform |
4 ≤ \(x\) ≤ 512 |
126 |
| regularization.user |
Float |
LogUniform |
1e-05 ≤ \(x\) ≤ 1 |
0.0389 |
| regularization.item |
Float |
LogUniform |
1e-05 ≤ \(x\) ≤ 1 |
0.146 |
| damping.user |
Float |
LogUniform |
1e-12 ≤ \(x\) ≤ 100 |
0.00633 |
| damping.item |
Float |
LogUniform |
1e-12 ≤ \(x\) ≤ 100 |
1.06e-11 |
| weight |
Float |
Uniform |
5 ≤ \(x\) ≤ 100 |
5.1 |
Final Result
Searching selected the following configuration:
{
'embedding_size': 126,
'regularization': {'user': 0.03891152508699394, 'item': 0.14615168345073604},
'damping': {'user': 0.006328080127355385, 'item': 1.0570018436811442e-11},
'weight': 5.099171622868312,
'epochs': 7
}
With these metrics:
{
'RBP': 0.17452988489343807,
'DCG': 11.273938090623373,
'NDCG': 0.41592125782256695,
'RecipRank': 0.33620098456541636,
'Hit10': 0.5872455902306648,
'max_epochs': 30,
'epoch_train_s': 10.926464226002281,
'epoch_measure_s': 87.06303000600019,
'done': True,
'training_iteration': 7,
'trial_id': '8dde8b04',
'date': '2025-07-29_04-21-22',
'timestamp': 1753777282,
'time_this_iter_s': 97.99300146102905,
'time_total_s': 810.4481289386749,
'pid': 273041,
'hostname': 'CCI-ws21',
'node_ip': '10.248.127.152',
'config': {
'embedding_size': 126,
'regularization': {'user': 0.03891152508699394, 'item': 0.14615168345073604},
'damping': {'user': 0.006328080127355385, 'item': 1.0570018436811442e-11},
'weight': 5.099171622868312,
'epochs': 7
},
'time_since_restore': 810.4481289386749,
'iterations_since_restore': 7
}
Parameter Analysis
Embedding Size
The embedding size is the hyperparameter that most affects the model’s fundamental logic, so let’s look at performance as a fufnction of it:
Iteration Completion
How many iterations, on average, did we complete?
How did the metric progress in the best result?
How did the metric progress in the longest results?
