EventKG+Click

EventKG+Click 是一个新颖的跨语言数据集，它反映了事件及其关系的语言特定相关性。该数据集旨在提供一个参考源来训练和评估以事件为中心的跨语言用户交互的新模型，特别关注知识图支持的模型。 EventKG+Click Dataset 基于两个数据源：

1. 反映真实世界用户与事件的交互及其在特定语言维基百科版本中的关系的维基百科点击流；和
2. EventKG 知识图谱，包含有关事件及其关系的语义信息，部分源自维基百科。
EventKG+Click 包含两个子集：

1.EventKG+Click_event，包含4000多个事件的相关分数、位置接近度、新近度和维基百科链接计数因子；和

2.EventKG+Click_relation 具有近万个以事件为中心的点击对，以及它们的语言特定点击次数、关系相关性和关系的共同提及，即在整个维基百科语言版本中提及两者来源的句子数量和目标。

EventKG+Click is a novel cross-lingual dataset that reflects the language-specific correlations of events and their relationships. This dataset is designed to serve as a reference resource for training and evaluating novel event-centric cross-lingual user interaction models, with a particular focus on knowledge graph-supported models. The EventKG+Click Dataset is built upon two data sources: 1. Wikipedia clickstream data that reflects real-world user interactions with events and their relationships across specific-language Wikipedia editions; and 2. the EventKG knowledge graph, which contains semantic information about events and their relationships, partially sourced from Wikipedia. EventKG+Click comprises two subsets: 1. EventKG+Click_event, which includes correlation scores, spatial proximity scores, recency, and Wikipedia link count factors for over 4,000 events; and 2. EventKG+Click_relation, which features nearly 10,000 event-centric click pairs, along with their language-specific click counts, relationship relevance, and co-mention counts of the relationship, i.e., the number of sentences across all Wikipedia language editions that mention both the source and target entities of the relationship.

搜集汇总

背景与挑战

背景概述

EventKG+Click是一个跨语言数据集，基于维基百科点击流和EventKG知识图谱，包含两个子集：EventKG+Click_event（4000多个事件的相关分数等）和EventKG+Click_relation（近万个以事件为中心的点击对）。该数据集旨在支持以事件为中心的跨语言用户交互模型的训练和评估。