IKNN Explicit

This page analyzes the hyperparameter tuning results for the ItemKNN scorer in explicit-feedback mode.

Parameter Search Space

Parameter Type Distribution Values Selected
max_nbrs Integer Uniform 2 ≤ \(x\) ≤ 50 32
min_nbrs Integer Uniform 1 ≤ \(x\) ≤ 5 3
min_sim Float LogUniform 1e-06 ≤ \(x\) ≤ 0.1 1.08e-05

Final Result

Searching selected the following configuration:

{'max_nbrs': 32, 'min_nbrs': 3, 'min_sim': 1.0797961971715764e-05}

With these metrics:

{
    'RBP': 0.03789515650431007,
    'NDCG': 0.3159376376957631,
    'RecipRank': 0.08845073915429313,
    'RMSE': 0.7588115820982672,
    'TrainTask': 'b21cbc66-6fc2-4c38-81ba-cddee540322d',
    'TrainTime': 2.8225089389998175,
    'TrainCPU': 2.8376400000000004,
    'TestTask': '54f9d97b-118e-41f6-ab55-d565332a85e4',
    'TestTime': 352.09064385700185,
    'TestCPU': 423.85758699999997,
    'timestamp': 1743040332,
    'checkpoint_dir_name': None,
    'done': True,
    'training_iteration': 1,
    'trial_id': '4191e_00014',
    'date': '2025-03-26_21-52-12',
    'time_this_iter_s': 358.84212160110474,
    'time_total_s': 358.84212160110474,
    'pid': 484480,
    'hostname': 'CCI-ws21',
    'node_ip': '10.248.127.152',
    'config': {'max_nbrs': 32, 'min_nbrs': 3, 'min_sim': 1.0797961971715764e-05},
    'time_since_restore': 358.84212160110474,
    'iterations_since_restore': 1,
    'experiment_tag': '14_max_nbrs=32,min_nbrs=3,min_sim=0.0000'
}

Parameter Behavior

Neighborhood Size

The neighborhood size is key argument to consider, let’s look at RMSE with repspect to that:

Filtering

There are two filtering parameters we also test — minimum neighbors and mininmum simialrity.