Title: Trend sensing via Twitter

Authors: Yavuz Selim Yilmaz; Muhammed Fatih Bulut; Cuneyt Gurcan Akcora; Murat Ali Bayir; Murat Demirbas

Addresses: Computer Science and Engineering Department, University at Buffalo, SUNY, Buffalo, New York, 14260, USA ' Computer Science and Engineering Department, University at Buffalo, SUNY, Buffalo, New York, 14260, USA ' Dipartimento di Informatica e Comunicazione, Università degli Studi dell'Insubria, Varese 21100, Italy ' Computer Science and Engineering Department, University at Buffalo, SUNY, Buffalo, New York, 14260, USA ' Computer Science and Engineering Department, University at Buffalo, SUNY, Buffalo, New York, 14260, USA

Abstract: Due to its ever increasing popularity, Twitter has become a pervasive information outlet. In this paper, we present a passive sensing framework for identifying trends via Twitter. In our framework, we use a multi-dimensional corpus for fine-granularity sensing of trends, and employ both vector-space and set-space methods for achieving accuracy. We present two applications of our framework. The first one is sensing trends in public opinion by using an emotion-category corpus. The second application is sensing trends in location-types in a city by using a location-category corpus. Our experiments show that the proposed methods are able to determine changes in trends effectively in both application scenarios.

Keywords: trend sensing; opinion mining; city-wide sensing; Twitter; public opinion; location.

DOI: 10.1504/IJAHUC.2013.056271

International Journal of Ad Hoc and Ubiquitous Computing, 2013 Vol.14 No.1, pp.16 - 26

Received: 30 Mar 2012
Accepted: 22 May 2012

Published online: 06 Sep 2013 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article