A three-layer model on users' interests mining. (February 2018)

• Record Type: Journal Article
• Authors: Yang, Renfeng; Xie, Wenbo; Chen, Duanbing
• Is Part Of: Journal of information science. Volume 44:Number 1(2018)
• Page Start: 136
• Page End: 144
• Publication Date: 2018-02
• Subjects: Interests mining -- semantic structures -- three-layer model -- words extracting; Information science -- Periodicals; Information science; Periodicals; 020.5
• DOI: 10.1177/0165551517743645
• Languages: English
• ISSNs: 0165-5515

Abstract

With the advent of big data era, social media plays an important role in many areas such as security and finance. Researchers pay more attention on mining users' interests through the social media data. A three-layer model (TLM) based on keyword extracting is proposed to mine users' interests, which includes candidate words extracting, semantic structures analysing and interest words ranking. The TLM mainly focuses on both self-importance and semantic-importance of interest words. In addition, the TLM also considers the interest drifting to track long-term and short-term interests of users. Experiments conducted on 10 SINA Weibo datasets show that TLM is more efficient than existing methods to identify users' interests based on hit rate.