Please use this identifier to cite or link to this item:
Title: Detecting polarity and concepts in climate change tweets with senticNet
Authors: Kiran, Mac Milin
Keywords: Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
Issue Date: 2022
Publisher: Nanyang Technological University
Source: Kiran, M. M. (2022). Detecting polarity and concepts in climate change tweets with senticNet. Final Year Project (FYP), Nanyang Technological University, Singapore.
Project: SCSE21-0232
Abstract: The Intergovernmental Panel on Climate Change (IPCC) released its Working Group 1 report on Climate Change in August 2021, which has sparked worldwide debates on social media and online forums. Many have since taken to Twitter, one of the most popular social media platforms with 217 million active daily users [2], to express their sentiments - advocating for policy changes or expressing disbelief in the shocking results. The vast number of tweets gives us the opportunity of mining user sentiment about Climate Change, thereby helping decision-makers at the government level to send across the right messages to educate the public and raise awareness. One of the essential tasks in analysing user opinions is sentiment classification. Sentiment classification uses Natural Language Processing (NLP) to classify textual data as positive, negative, or neutral. Sentiment Analysis (SA) can be considered as a big ‘suitcase’ problem, which tackles multiple sub-problems of Natural Language Processing (NLP) like keyword extraction, polarity detection, and sarcasm detection, to name a few [3]. Most recent research in sentiment analysis focuses on subsymbolic AI, i.e., machine learning, a powerful way to analyse large amounts of data, categorizing and classifying. It is essential to integrate logical reasoning to detect meaningful patterns in natural language text and statistical and vector categorizations. Thus, this paper makes use of SenticNet (specifically SenticNet 6), a knowledge base that makes use of symbolic and subsymbolic AI to increase the accuracy of NLP [3] significantly. While climate change opinions have been mined before for the 2013 IPCC Working Group 1 Report, the analysis focused only on statistical methods, i.e., subsymbolic AI. This project covers climate change opinions after the 2021 IPCC Working Group 1 report was published and uses symbolic and subsymbolic AI to classify sentiments.
Schools: School of Computer Science and Engineering 
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
  Restricted Access
1.17 MBAdobe PDFView/Open

Page view(s)

Updated on Sep 22, 2023

Download(s) 50

Updated on Sep 22, 2023

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.