Microblog search and filtering with real-time dynamics based on BM25

Wei Gao, Zhongyu Wei, Kam Fai Wong

Research output: Chapter in Book/Report/Conference proceedingChapter


Microblogs such as Twitter are considered faster first-hand sources of information with many real-time fashions. We report our work in the real-time ad hoc search and filtering tasks of TREC 2012 microblog track. Our system is built based on the traditional BM25 relevance model, in which specific techniques are tried out to respond to the need of finding relevant tweets. In the real-time ad hoc task, we applied a peak detection algorithm for the process of blind feedback. We also tried to automatically combine the search results of multiple retrieval techniques. In the real-time filtering pilot task, we examine the effectiveness of some typical filtering methods previously used in TREC filtering track.

Original languageEnglish
Title of host publicationSocial Media Content Analysis
Subtitle of host publicationNatural Language Processing and Beyond
PublisherWorld Scientific Publishing Co. Pte Ltd
Number of pages12
ISBN (Electronic)9789813223615
ISBN (Print)9789813223608
Publication statusPublished - 1 Jan 2017


ASJC Scopus subject areas

  • Computer Science(all)

Cite this

Gao, W., Wei, Z., & Wong, K. F. (2017). Microblog search and filtering with real-time dynamics based on BM25. In Social Media Content Analysis: Natural Language Processing and Beyond (pp. 19-30). World Scientific Publishing Co. Pte Ltd. https://doi.org/10.1142/9789813223615_0002