arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2603.01299physics.soc-ph
收藏

不同语言中通过维基百科编辑呈现的偏好时间模式

Temporal patterns of preferences through Wikipedia editing in different languages

Valencian International University(瓦伦西亚国际大学);Autonomous University of Madrid(马德里自治大学)

David André Villamil Carrillo, Yérali Gandica

AI总结:

研究通过分析维基百科编辑数据,揭示不同语言社区的昼夜编辑节律差异,强调时间维度在跨文化计算分析中的重要性。

摘要:

维基百科上的时间编辑模式为探索语言社区中的文化动态提供了独特的计算视角。本研究分析了2001-2010年间十一种语言版本的编辑活动,代表了多样化的语言和文化社区。我们应用层次聚类与通过PCA和自编码器进行的降维技术,对集体行为的静态(分类)和时间维度进行处理。结果表明,语言社区表现出受文化和社会因素影响的不同的昼夜编辑节律。关键的是,静态和时间聚类产生了显著不同的社区分组,证明了时间是跨文化计算分析中至关重要的——但往往被忽视——维度。这些发现有助于我们理解文化身份如何在大规模数字痕迹数据中表现出来,并为未来利用在线平台作为集体文化行为代理的研究提供方法论启示。

英文摘要:

Temporal editing patterns on Wikipedia provide a computational lens on cultural dynamics across linguistic communities. We analyse a decade of editorial activity (2001--2010) in eleven Wikipedia language editions and apply hierarchical clustering to two representations of collective behaviour: a static one, based on the thematic composition of each edition, and a temporal one, based on its weekly activity cycle within each thematic category. For the temporal representation we compare two normalisations: one retains only the shape of the weekly cycle, and the other combines it with thematic composition. The linguistic communities exhibit distinct circadian and weekly editing rhythms. Some groupings are recovered by both representations and by both activity metrics (edits and editors), most notably Russian and Chinese, whereas others, such as German and French, emerge only from temporal activity rhythms. Temporal information thus reveals structure that thematic composition alone does not, which points to time as a relevant and often neglected dimension in cross-cultural analyses of digital trace data, and in studies that use online platforms as proxies for collective cultural behaviour.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑