WEBQUESTION数据集

承担单位北京大学在项目执行过程中研究了基于知识图谱的问答算法，使用WEBQUESTION数据集进行关键指标的评测。该数据集包含6642个问题-答案对，通过Google Suggest API爬取问题，再人工标注答案得到，使用Freebase作为知识图谱，所有答案均为知识图谱中的实体。数据集存储规模约1MB。

Peking University, the undertaking institution of this project, conducted research on knowledge graph-based question answering algorithms during the project implementation. The WEBQUESTION dataset was used to evaluate key performance metrics. This dataset contains 6,642 question-answer pairs, which were developed by crawling questions via the Google Suggest API and then manually annotating their corresponding answers. Freebase was employed as the knowledge graph, and all answers are entities within this knowledge graph. The storage size of the dataset is approximately 1 MB.

搜集汇总

背景与挑战

背景概述

WEBQUESTION数据集是一个用于知识图谱问答算法评测的基准数据集，包含6642个通过Google Suggest API爬取并人工标注的问题-答案对，所有答案均基于Freebase知识图谱中的实体。该数据集规模较小（约1MB），适用于自然语言处理和人工智能领域的研究，特别是问答系统开发与评估。