CFKG

Reference:

Qingyao Ai et al. “Learning heterogeneous knowledge base embeddings for explainable recommendation.” in MDPI 2018.

class recbole.model.knowledge_aware_recommender.cfkg.CFKG(config, dataset)

Bases: recbole.model.abstract_recommender.KnowledgeRecommender

CFKG is a knowledge-based recommendation model, it combines knowledge graph and the user-item interaction graph to a new graph. In this graph, user, item and related attribute are viewed as entities, and the interaction between user and item and the link between item and attribute are viewed as relations. It define a new score function as follows:

\[d (u_i + r_{buy}, v_j)\]

Note

In the original paper, CFKG puts recommender data (u-i interaction) and knowledge data (h-r-t) together for sampling and mix them for training. In this version, we sample recommender data and knowledge data separately, and put them together for training.

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(user, item)

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.

input_type = 2

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

training: bool

class recbole.model.knowledge_aware_recommender.cfkg.InnerProductLoss

Bases: torch.nn.modules.module.Module

This is the inner-product loss used in CFKG for optimization.

forward(anchor, positive, negative)

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.

training: bool