图书情报知识 ›› 2018, Vol. 0 ›› Issue (1): 27-35.doi: 10.13366/j.dik.2018.01.027

• 博士论坛 • 上一篇    下一篇

跨设备搜索引擎结果页面注意力分布研究——基于眼动视觉数据的实证分析

梁少博,吴丹,董晶,唐源   

  • 出版日期:2018-01-10 发布日期:2018-01-10

A Research on the Distribution of Attention in Cross-device Search Engine Result Page: An Empirical Study Based on Eye Tracking Data Analysis

  • Online:2018-01-10 Published:2018-01-10

摘要:

用户在搜索引擎结果页面的视觉注视行为,一直是信息检索领域的重要研究内容,有助于优化搜索引擎结果页面(Serach Engine Result Page, SERP)的布局,提升用户搜索效率。而针对用户在跨设备搜索情境下的SERP注视行为的研究还较少。本研究通过跨设备搜索实验,对用户在不同跨设备情境下的SERP视觉行为分布展开研究。研究发现,用户在跨设备后,其视觉注意力相比之前有所分散,关注点减少。跨设备后,用户的“眼动熵”值在SERP的搜索结果列表中,呈现出总体上升的趋势。用户在跨设备后对SERP首屏的搜索结果区域内搜索结果摘要的关注度最高,对于记录跨设备历史信息的区域关注度提升最高,这说明搜索引擎为用户提供的跨设备历史信息能够有效地帮助用户恢复搜索任务,提高用户的搜索效率。在单条搜索结果区域内,跨设备前后用户的视觉分布不存在显著性差异。

关键词: 跨设备搜索, 搜索行为, 眼动追踪, 用户研究, 搜索引擎结果页面

Abstract:

Users’ visual gaze behaviors in search engine result page (SERP) has been an important research content in information retrieval field, which can help optimize the layout of SERP and enhance users’ search efficiency. However, few researches have been done on the gaze behavior in SERP of cross device search. Through a crossdevice search experiment, this paper studies the distribution of user’s visual behavior in SERP in different inter device context. The results reveals  that users’ visual attention isless focused comparing the first device. The value of “Eyetracking entropy” in result list of SERP shows a upward trend generally after device transition. Users pay the most attention to the search results snippets in the first screen of SERP after the device transition, and the highest promotion of visual attention occurs in the region about crossdevice history. This indicates that the crossdevice history information provided by the search engine can effectively help users recover the search task and improve the search efficiency. In the unique search results region, there is no significant difference in the visual attention.

Keywords: Cross-device search, Search behavior, Eye tracking, User study, Search engine result page(SERP)