RC‐MF

Running RC-MF]

Instructions for:

• running RC-MF with default parameters;
• running RC-MF with Bayesian optimization;
• selecting datasets and random seeds;
• evaluating residual diagnostics;
• locating generated result files.

Main experiment command

Experiments are executed through:

python experiment_runner.py --dataset DATASET_NAME

By default, the runner evaluates RC-MF and all registered Cornac baselines.

To run only RC-MF:

python experiment_runner.py \
    --dataset Beauty \
    --epochs 100 \
    --seed 0 \
    --models rc_mf

To run RC-MF with Bayesian optimization:

python experiment_runner.py \
    --dataset Beauty \
    --epochs 100 \
    --seed 0 \
    --use_bo \
    --models rc_mf

Supported evaluation metrics

• RMSE;
• MSE;
• MAE;
• (R^2).

It also reports top-(K) ranking metrics when the model supports candidate-item scoring:

• Precision@5, @10, and @20;
• Recall@5, @10, and @20;
• NDCG@5, @10, and @20.

Configuration

Model defaults, hyperparameter search spaces, and Bayesian-optimization budgets are defined in:

config/model_search_spaces.yml