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 4
min_sim Float LogUniform 1e-06 ≤ \(x\) ≤ 0.1 2.62e-05

Final Result

Searching selected the following configuration:

{'max_nbrs': 32, 'min_nbrs': 4, 'min_sim': 2.6212738427918034e-05}

With these metrics:

{
    'RBP': 0.006438200633294398,
    'LogRBP': -1.2512662129028485,
    'NDCG': 0.28539924916136455,
    'RecipRank': 0.02571444198197259,
    'RMSE': 0.7938979832134402,
    'TrainTask': '51dcbe4c-d2ef-4239-9b04-aaea118dd247',
    'TrainTime': 5.074715482012834,
    'TrainCPU': 5.160131,
    'TestTask': '0fbfc962-f683-41d3-bd15-d1f749203417',
    'TestTime': 347.48131107294466,
    'TestCPU': 2159.618178,
    'timestamp': 1746673103,
    'checkpoint_dir_name': None,
    'done': True,
    'training_iteration': 1,
    'trial_id': 'd196222f',
    'date': '2025-05-07_22-58-23',
    'time_this_iter_s': 365.164648771286,
    'time_total_s': 365.164648771286,
    'pid': 1586813,
    'hostname': 'gracehopper1',
    'node_ip': '192.168.225.60',
    'config': {'max_nbrs': 32, 'min_nbrs': 4, 'min_sim': 2.6212738427918034e-05},
    'time_since_restore': 365.164648771286,
    'iterations_since_restore': 1,
    'experiment_tag': '29_max_nbrs=32,min_nbrs=4,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.