SHAN

Reference:

Ying, H et al. “Sequential Recommender System based on Hierarchical Attention Network.”in IJCAI 2018

class recbole.model.sequential_recommender.shan.SHAN(config, dataset)

Bases: recbole.model.abstract_recommender.SequentialRecommender

SHAN exploit the Hierarchical Attention Network to get the long-short term preference first get the long term purpose and then fuse the long-term with recent items to get long-short term purpose

calculate_loss(interaction)

Calculate the training loss for a batch data.

Parameters

interaction (Interaction) – Interaction class of the batch.

Returns

Training loss, shape: []

Return type

torch.Tensor

forward(seq_item, user, seq_item_len)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

full_sort_predict(interaction)

full sort prediction function. Given users, calculate the scores between users and all candidate items.

Parameters

interaction (Interaction) – Interaction class of the batch.

Returns

Predicted scores for given users and all candidate items, shape: [n_batch_users * n_candidate_items]

Return type

torch.Tensor

init_weights(module)

inverse_seq_item(seq_item, seq_item_len)

inverse the seq_item, like this

[1,2,3,0,0,0,0] – after inverse –>> [0,0,0,0,1,2,3]

long_and_short_term_attention_based_pooling_layer(long_short_item_embedding, user_embedding, mask=None)

fusing the long term purpose with the short-term preference

long_term_attention_based_pooling_layer(seq_item_embedding, user_embedding, mask=None)

get the long term purpose of user

predict(interaction)

Predict the scores between users and items.

Parameters

interaction (Interaction) – Interaction class of the batch.

Returns

Predicted scores for given users and items, shape: [batch_size]

Return type

torch.Tensor

reg_loss(user_embedding, item_embedding)

training: bool