Pairwise Model

Model using pairwise learning to rank.

src.models.pairwise.bubble_sort(pairwise_results, documents) → list

.

Parameters

• pairwise_results (list) –

• documents (list) –

Return type

documents (list)

src.models.pairwise.create_dataloader(X, y, batch_size: int = 50) → torch.utils.data.dataloader.DataLoader

.

Parameters

• () (y) –

• () –

• batch_size (int) –

src.models.pairwise.create_test_combinations(top: pandas.core.frame.DataFrame, k: int = 50) → tuple

Creates test combinations.

Parameters

• top (pd.DataFrame) –

• k (int) –

Returns

X_irrelevant_test ():

Return type

X_relevant_test ()

src.models.pairwise.pairwise_optimize(model, results: pandas.core.frame.DataFrame, X, y, X_test, top_k: int = 50, train: bool = True) → pandas.core.frame.DataFrame

.

Parameters

• () (X_test) –

• results (pd.DataFrame) –

• () –

• y (pd.DataFrame) –

• () –

• top_k (int) –

• train (Boolean) –

Return type

results (pd.DataFrame)

src.models.pairwise.train_pairwise(network, X, y, num_epochs: int = 10)

.

Parameters

• () (network) –

• X (int) –

• y (int) –

• num_epochs (int) –