Bias Predictor

This page analyzes the behavior and hyperparameter search for the Bias scoring model on ML100K.

Parameter Search Space

Parameter	Type	Distribution	Values	Selected
damping.user	Float	LogUniform	0.1 ≤ \(x\) ≤ 100	5.63
damping.item	Float	LogUniform	0.1 ≤ \(x\) ≤ 100	10.2

Final Result

Searching selected the following configuration:

{'damping': {'user': 5.628454382751774, 'item': 10.215483921135197}}

With these metrics:

{
    'RBP': 0.021412473919955367,
    'LogRBP': -0.04954166354311518,
    'NDCG': 0.2033253680315756,
    'RecipRank': 0.08062503023343114,
    'RMSE': 0.9582030959624462,
    'TrainTask': 'dcf3cd54-ba17-4505-93d8-99edd3a834b3',
    'TrainTime': 0.04790819197660312,
    'TrainCPU': 0.0517690000000004,
    'TestTask': '6b1e8a5a-d8fb-4cf0-a24e-c7bccc59cb6c',
    'TestTime': 0.3331394880078733,
    'TestCPU': 0.3347000000000002,
    'timestamp': 1746582329,
    'checkpoint_dir_name': None,
    'done': True,
    'training_iteration': 1,
    'trial_id': '765fe947',
    'date': '2025-05-06_21-45-29',
    'time_this_iter_s': 0.5471127033233643,
    'time_total_s': 0.5471127033233643,
    'pid': 671507,
    'hostname': 'gracehopper1',
    'node_ip': '192.168.225.60',
    'config': {'damping': {'user': 5.628454382751774, 'item': 10.215483921135197}},
    'time_since_restore': 0.5471127033233643,
    'iterations_since_restore': 1,
    'experiment_tag': '100_item=10.2155,user=5.6285'
}

Metric Response

How does RMSE change with each setting independently?

User Damping

Item Damping

Best Configurations

Since this is an explicit-feedback rating prediction model, our primary search criteria is RMSE. The configuration with the best RMSE is:

	config.damping.user	config.damping.item	RBP	RMSE
Method
Random	5.628454	10.215484	0.021412	0.958203

If we instead searched for RBP, we would select:

	config.damping.user	config.damping.item	RBP	RMSE
Method
Random	6.522858	94.023965	0.034546	0.970885

Search Geometry

What is the geometry of the search space?