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	5.16
damping.item	Float	LogUniform	0.1 ≤ \(x\) ≤ 100	9.91

Final Result

Searching selected the following configuration:

{'damping': {'user': 5.161989087913108, 'item': 9.905192700858715}}

With these metrics:

{
    'RBP': 0.021345765895989667,
    'LogRBP': -0.05266190789250125,
    'NDCG': 0.2032090004834285,
    'RecipRank': 0.08046790552853415,
    'RMSE': 0.958215342391105,
    'TrainTask': '1fec5c09-cd2e-48ba-b175-c3977b9cf512',
    'TrainTime': 0.04613151994999498,
    'TrainCPU': 0.047717000000000065,
    'TestTask': 'e6911bfb-c80a-469a-afb6-97871907919b',
    'TestTime': 0.393404096015729,
    'TestCPU': 0.39629800000000004,
    'timestamp': 1746581134,
    'checkpoint_dir_name': None,
    'done': True,
    'training_iteration': 1,
    'trial_id': '1d8d2_00090',
    'date': '2025-05-06_21-25-34',
    'time_this_iter_s': 0.5912454128265381,
    'time_total_s': 0.5912454128265381,
    'pid': 594651,
    'hostname': 'gracehopper1',
    'node_ip': '192.168.225.60',
    'config': {'damping': {'user': 5.161989087913108, 'item': 9.905192700858715}},
    'time_since_restore': 0.5912454128265381,
    'iterations_since_restore': 1,
    'experiment_tag': '90_item=9.9052,user=5.1620'
}

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	5.161989	9.905193	0.021346	0.958215

If we instead searched for RBP, we would select:

	config.damping.user	config.damping.item	RBP	RMSE
Method
Random	5.125235	96.629343	0.035106	0.971116

Search Geometry

What is the geometry of the search space?