基于LDA的群组聊天行为研究

发布时间：2018-03-17 05:15

本文选题：LDA模型　切入点：主题提取　出处：《情报科学》2017年12期 　论文类型：期刊论文

【摘要】：【目的/意义】随着社交网络的普及与快速发展,人们越来越多地依赖于网络聊天工具进行交流,针对QQ群组聊天信息过载用户无法从聊天记录中快速获取所需信息的问题,本文开展了聊天热点主题提取和QQ群组用户聊天行为分析的研究。【方法/过程】采集了一个技术类QQ群的聊天数据,利用Gibbs算法和LDA模型提取群组聊天数据中的主题并对其进行分析。【结果/结论】发现群组的主题可以分为三类:技术类、生活类和综合类。其中,技术类话题讨论的高峰集中在工作时间,没有继承性;大家普遍关心生活类话题,该话题有继承性。由于群组聊天的即时性、交互性和网络领袖的影响,一个时间段内群中只有一个热点主题。该研究结果可为群组聊天行为和热点分析提供参考。
[Abstract]:[purpose / meaning] with the popularity and rapid development of social networks, people rely more and more on network chat tools for communication, aiming at the problem that QQgroup chat information overload users can not quickly get the required information from chat records. In this paper, the topic extraction of chat hotspots and the analysis of chat behavior of QQgroup users are studied. [method / process] A chat data of a technical QQgroup is collected. Using Gibbs algorithm and LDA model, the topic of group chat data is extracted and analyzed. [results / conclusions] it is found that the topic of group can be divided into three categories: technology, life and synthesis. The peak of the discussion on the technology topic is focused on the working hours and has no inheritance. There is a general concern about the topic of life, which has inheritance. Because of the immediacy, interactivity of group chat and the influence of network leaders, There is only one hot topic in a group within a period of time. The results can be used as a reference for group chat behavior and hot spot analysis.
【作者单位】：长春理工大学计算机科学与技术学院;东北师范大学马克思主义学院;
【基金】：国家社科基金(17BSH135)
【分类号】：G252;TP391.1

【相似文献】