关键词: 网络百科, 用户兴趣, 动态变化, 时间序列, 聚类分析

Abstract:

Understanding users deeply is the foundation of optimizing the wiki systems. By introducing the concept of outstanding interest and stable interest, this article uses some methods such as time series clustering to explore the dynamic law of users’ interests based on about five hundred thousand data records of Baidu Baike in six month in accordance with the disadvantages of users’ interests. This research finds that the users’ interests are stable or accidental, and there are significant differences among them. Most users make contributions just for some accidental reasons. There is a shortage of experts who can keep contributing on one topic over long periods. This may be one reason why the quality of wiki articles is widely low. Therefore, it is necessary to attract experts to contributing for wikis while continuing to motivate normal users. To ensure the effect of personalized recommendation, users’ interests must be acquired from multiple sources.

Key words: Online encyclopedia；Users&rsquo, interests；Dynamic change；Time series；Clustering analysis