Please use this identifier to cite or link to this item:
Title: FIRE : Finding Important News REports
Authors: Mamo, Nicholas
Keywords: Information retrieval -- Computer programs
Expert systems (Computer science)
Information filtering systems
Issue Date: 2017
Abstract: Every day, an immeasurable number of news items are published. Social media greatly contributes to the dissemination of information, and as a result, at any point in time, it is impossible to stay on top of every article. Among the most popular social networks, Twitter stands out for a variety of reasons. Twitter is a valuable source of information mainly due to its large user base and the immediateness with which news is reported and spread. To this end, the proposed solution - Finding Important News Reports (FIRE) - aims to exploit the information and metrics available on Twitter to identify and track breaking news, and the defining articles that discuss them. More specifically, FIRE examines how the different metrics available on Twitter, such as its users' engagement with tweets, can help detect breaking news. The developed solution is split into different sections, namely clustering, emerging topic detection and tracking, spam and noise removal, and finally the identification of related news reports. Each part has a clearly-defined role in the bigger picture, and has been the subject of research in this project. These different methods, albeit widely-researched in information systems, present context-specific problems when dealing with the micromessages of Twitter. FIRE demonstrates how even a small sample of tweets can be used to extract newsworthy stories that the general population deems important. In fact, Twitter's conversation habits do nothing to shackle the clustering process. Furthermore, it is shown that the adopted workflow can pick up emerging topics earlier than Twitter itself, and goes a step farther by capturing stories that are not detected by the social network. Nonetheless, the results obtained in this project indicate the need for a reliable and efficient spam and noise filtering tool, and the dependency of the results on such a tool.
Description: B.SC.IT(HONS)
Appears in Collections:Dissertations - FacICT - 2017
Dissertations - FacICTAI - 2017

Files in This Item:
File Description SizeFormat 
  Restricted Access
2.48 MBAdobe PDFView/Open Request a copy

Items in OAR@UM are protected by copyright, with all rights reserved, unless otherwise indicated.