Fig 1.
A subgraph of Fig 4.
Fig 2.
1-N relation in knowledge graph.
Fig 3.
The transition probability process of Node2Vec.
Fig 4.
A knowledge graph of user-movies.
Fig 5.
A simple knowledge graph of user-movies1.
Fig 6.
Reconstructed knowledge graph.
Fig 7.
Entity node transition probability.
Fig 8.
Relation node transition probability.
Fig 9.
The KG2Vec model (a. the prediction of the relation using the CBOW, b. the prediction of the entity using the CBOW, c. the prediction of the entity’s context using Skip-gram, and d. the prediction of the relation’s context with Skip-gram).
Table 1.
Data sets.
Fig 10.
Line chart of P (a is the entity data of the FB15K data set, b is the relational data of the FB15K data set, c is the entity data of the WN18 data set, and d is the parameter relational data of the WN18 data set).
Fig 11.
Line chart of q (a is the entity data of the FB15K data set, b is the relational data of the FB15K data set, c is the entity data of the WN18 data set, and d is the parameter relational data of the WN18 data set).
Table 2.
Experimental results of KG2Vec model on FB15K-237.
Table 3.
Experimental results of KG2Vec model on WN18.