Questions are content: A taxonomy of questions in a microblogging environment

Abstract

Microblogging services such as twitter.com have become popular venues for informal information interactions. An important aspect of these interaction is question asking. In this paper we report results from an analysis of a large sample of data from Twitter. Our analysis focused on the characteristics and strategies that people bring to asking questions in microblogs. In particular, based on our analysis, we propose a taxonomy of questions asked in microblogs. We find that microblog authors express questions to accomplish a wide variety of social and informational tasks. Some microblog questions seek immediate answers, while others accrue information over time. Our overarching finding is that question asking in microblogs is strongly tied to peoples' naturalistic interactions, and that the act of asking questions in Twitter is not analogous to information seeking in more traditional information retrieval environments.

Estimation methods for ranking recent information

“…Temporal factors have been shown to provide leverage in several IR-related problems, including document ranking [6] and query difficulty prediction [8].…”

“…Problems such as collaborative and social search [9,11] draw attention to the role that relationships between people play in a variety of search settings. Other work pays specific attention to the social aspects of information seeking in microblog settings [8,16]. More traditional treatments of IR in microblogs is offered in [2,7,10] and [23].…”

Question identification on twitter

“…Interrogative tweets detection is closely related to question (subquestion) finding in UGC such as forums, community question answering (CQA) portals and microblogs. There are two divisions for question finding in UGC: rulebased approach [2] and learning-based approach [1,6]. Rulebased approach usually designs several rules from heuristics or observations to check whether a thread or tweet is question or not, while learning-based approach constructs a binary classifier with lexical and/or syntactic features.…”

“…and Winget [2] found that 13% of their randomly sampled 2-million tweets corpus were questions. According to the blog of Twitter, on average 1,620 tweets were posted every second in March 2011…”

“…According to the taxonomy in [2], many questions are not requesting information but providing information, like suggesting a Q&A pair or expressing an opinion. Different from the meanings of questions in [2] and [3], we restrict the scope of questions to be those tweets which require some information or help and thus need to be answered. In the following, we use the term "qweets" to represent these tweets.…”