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 34
min_nbrs Integer Uniform 1 ≤ \(x\) ≤ 5 1
min_sim Float LogUniform 1e-06 ≤ \(x\) ≤ 0.1 7.9e-05

Final Result

Searching selected the following configuration:

{'max_nbrs': 34, 'min_nbrs': 1, 'min_sim': 7.896631856539726e-05}

With these metrics:

{
    'RBP': 0.06458015967040781,
    'NDCG': 0.2893437492019393,
    'RecipRank': 0.11442041001021389,
    'RMSE': 0.7772569359349784,
    'TrainTask': '4dd84a2b-f56c-40b8-916d-63316baa1334',
    'TrainTime': 2.7228366880008252,
    'TrainCPU': 2.7229550000000007,
    'TestTask': '63825108-b8d4-4fd7-8c3a-53a45359c25b',
    'TestTime': 279.60471785999835,
    'TestCPU': 280.504917,
    'timestamp': 1746062037,
    'checkpoint_dir_name': None,
    'done': True,
    'training_iteration': 1,
    'trial_id': '8166a5c9',
    'date': '2025-04-30_21-13-57',
    'time_this_iter_s': 287.81806921958923,
    'time_total_s': 287.81806921958923,
    'pid': 339262,
    'hostname': 'CCI-ws21',
    'node_ip': '10.248.127.152',
    'config': {'max_nbrs': 34, 'min_nbrs': 1, 'min_sim': 7.896631856539726e-05},
    'time_since_restore': 287.81806921958923,
    'iterations_since_restore': 1,
    'experiment_tag': '70_max_nbrs=34,min_nbrs=1,min_sim=0.0001'
}

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.