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 24
min_nbrs Integer Uniform 1 ≤ \(x\) ≤ 5 4
min_sim Float LogUniform 1e-06 ≤ \(x\) ≤ 0.1 0.0101

Final Result

Searching selected the following configuration:

{'max_nbrs': 24, 'min_nbrs': 4, 'min_sim': 0.01011912252100525}

With these metrics:

{
    'RBP': 0.007097633001586398,
    'LogRBP': -1.1537539608048446,
    'NDCG': 0.2498489687301457,
    'RecipRank': 0.026403775630755574,
    'RMSE': 0.8156804628304165,
    'TrainTask': '5100ab14-6bf6-45a7-a404-0ca5b169d519',
    'TrainTime': 7.264380648965016,
    'TrainCPU': 7.390893999999999,
    'TestTask': 'c7ff5cc2-24f3-4ab4-9239-63f11c5f7db3',
    'TestTime': 557.2476165389526,
    'TestCPU': 3211.310595,
    'timestamp': 1746669674,
    'checkpoint_dir_name': None,
    'done': True,
    'training_iteration': 1,
    'trial_id': 'edb3096a',
    'date': '2025-05-07_22-01-14',
    'time_this_iter_s': 590.3438272476196,
    'time_total_s': 590.3438272476196,
    'pid': 1562854,
    'hostname': 'gracehopper1',
    'node_ip': '192.168.225.60',
    'config': {'max_nbrs': 24, 'min_nbrs': 4, 'min_sim': 0.01011912252100525},
    'time_since_restore': 590.3438272476196,
    'iterations_since_restore': 1,
    'experiment_tag': '17_max_nbrs=24,min_nbrs=4,min_sim=0.0101'
}

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.