Detecting Points of View from News Articles.

Date of Submission

December 2015

Date of Award

Winter 12-12-2016

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science


Computer Vision and Pattern Recognition Unit (CVPR-Kolkata)


Majumdar, Debapriyo (CVPR-Kolkata; ISI)

Abstract (Summary of the Work)

Conventional models concentrate on what the journalist perceives as news. But the news process is a two-way transaction, involving both news producer (the journalist) and the news receiver (the audience), although boundary between the two is rapidly blurring with the growth of citizen journalism and interactive media. Different newspapers have different perspectives from different journalists regarding a topic. In this thesis, we propose a novel approach to automatically extracts the multiple Points of View from different newspapers articles talking on a similar topic. Rather than ranking or summarisation of cluster topics, we try to bridge the gap of information a audience might have when we doesn’t read multiple newspaper. Thus we can view the documents as being composed of different Points of View regarding a same event clustered by source and target, which we have to infer, and the visible variables which are the words of documents are just means of expressing these views and the weighted links of the graph defines the sentiment polarity, which is the data that we have. A sentence quotation can be viewed as describing a single event or maybe connecting different events of the document. Here we are dealing with extraction based on these opinion and identifying the polarity of these views of the multiple documents and giving the user Information dessert in form of Points of View.Our approach is distinguished from existing approaches in that we use models to capture the Points of View after identification of important entities [Source,Target]. Work also involves picking up the sentences without paying attention to the details of grammar and structure of the documents and also stating the polarity of the respective views.


ProQuest Collection ID:

Control Number


Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.


This document is currently not available here.