Abstract:With the rapid development of computer technology,the access to educational resources has become more and more convenient. Massive learning resources make it difficult for learners to reasonably organize the learning order of knowledge concepts. When students learn new knowledge concepts, they are often related to previously learned knowledge, which are generally called prerequisite knowledge concept of the new knowledge concept. Moreover, the similarity between knowledge concepts is also important. If a knowledge concept and a new knowledge concept have similar learning contents, for example, the message queue framework RocketMQ and Kafka have similar underlying principles, we think they have similar relation. However, which is usually ignored in existing researches. Therefore, previous methods only focus on identifying if there is a relationship between two knowledge concepts, instead of judging what kind of relationship. This paper proposes a new link prediction model based on representation learning, namely LPRL-GCNN model - LPRL-GCNN (Link Prediction model based on Representation Learning). This model can not only predict new knowledge concept associations, but also predict the relationship types between knowledge concepts, that is, multi relationship link prediction tasks. We aggregate the adjacent features of nodes through RGCN, obtain the edge features between the two nodes through convolution, classify according to the edge features, and complete the multi relationship link prediction task. Our experiments on two datasets show that the proposed method achieves the advanced results compared with existing methods.

A Multi-Hop Link Prediction Approach Based on Reinforcement Learning in Knowledge Graphs.

Cross-KG Link Prediction by Learning Substructural Semantics

RLPath: a knowledge graph link prediction method using reinforcement learning based attentive relation path searching and representation learning

Meta Relational Learning for Few-Shot Link Prediction in Knowledge Graphs

Generalized Relation Learning with Semantic Correlation Awareness for Link Prediction

Location Based Link Prediction for Knowledge Graph

Link Prediction using Embedded Knowledge Graphs

Dynamic relation learning for link prediction in knowledge hypergraphs

Knowledge Graph Large Language Model (KG-LLM) for Link Prediction

Geography-Enhanced Link Prediction Framework for Knowledge Graph Completion.

Knowledge Graph Link Prediction Based on Subgraph Reasoning

Multi-hop Knowledge Reasoning with Deep Reinforcement Learning

LPRL-GCNN for Multi-relation Link Prediction in Education

Attention-based Exploitation and Exploration Strategy for Multi-Hop Knowledge Graph Reasoning

Multi-level Recommendation Reasoning over Knowledge Graphs with Reinforcement Learning

Temporal link prediction via reinforcement learning

Enhancing Multi-Hop Knowledge Graph Reasoning through Reward Shaping Techniques

Multiple Run Ensemble Learning with Low-Dimensional Knowledge Graph Embeddings

Multi-hop Knowledge Graph Reasoning Based on Hyperbolic Knowledge Graph Embedding and Reinforcement Learning

Convolutional Models with Multi-Feature Fusion for Effective Link Prediction in Knowledge Graph Embedding

A Survey on Knowledge Graph Embeddings for Link Prediction