FB15K-237

FB15K-237数据集是基于Freebase知识库构建的一个广泛使用的知识图谱数据集，主要用于知识图谱补全和链接预测任务。它包含14,541个实体和237种关系类型，共有310,079个三元组，其中训练集包含272,115个三元组，验证集包含17,535个三元组，测试集包含20,466个三元组。该数据集去除了逆向关系，减少了数据冗余，提高了模型的训练效率。它在知识图谱研究中具有重要的地位，尤其在关系推理和链接预测任务上被广泛应用。

The FB15K-237 dataset is a widely used knowledge graph dataset constructed based on the Freebase knowledge base, primarily intended for knowledge graph completion and link prediction tasks. It contains 14,541 entities and 237 relationship types, with a total of 310,079 triples. Specifically, the training set includes 272,115 triples, the validation set contains 17,535 triples, and the test set has 20,466 triples. This dataset removes inverse relations to reduce data redundancy and improve the training efficiency of models. It holds an important position in knowledge graph research and is widely applied, especially in relational reasoning and link prediction tasks.

搜集汇总

背景与挑战

背景概述

FB15K-237是基于Freebase知识库构建的知识图谱数据集，包含14,541个实体和237种关系类型，共310,079个三元组，主要用于知识图谱补全和链接预测任务。该数据集去除了逆向关系，减少了冗余，提高了训练效率，在知识图谱研究中具有重要地位。