Semi-Supervised Classification of Climate Change & Population Related Tweets based on Hash Tags and Accounts annotation

Chongtham Rajen Singh; Dr. R. Gobinath; Dr. Nongmaithem Ajith Singh

Semi-Supervised Classification of Climate Change & Population Related Tweets based on Hash Tags and Accounts annotation

Authors: Chongtham Rajen Singh, Dr. R. Gobinath, Dr. Nongmaithem Ajith Singh

Unique Paper ID: 156901
Volume: 9
Issue: 5
PageNo: 294-302

Keywords: Climate Population Denier Believer Annotation Hash Tag Account Vectorizer Supervised Classification

Abstract:
Many researchers work on climate change related tweets to predict and determine whether climate change is real or hoax based on sentiment analysis using labeled dataset. Others work on predefined climate change denier hash tags or denier twitter accounts to annotate these tweets. This paper illustrates the combine usages of denier hash tags and predefined twitter accounts that are primarily denier of climate change propaganda in Twitter. Denier accounts and hash tags collected from across various papers, articles are used along with additional wild character search techniques of seed phrases related to climate change and population growth related words or phrases together and annotated the unlabeled dataset extracted from Twitter. It is an automatic annotation technique. This annotated subset of data is used to train baseline Supervised Classification Models in combination with two types of frequency-based word vectorizers to analyze the performance measure of each model with different feature variation of n-gram combination. As per the analysis Linear Support Vector Machine algorithm along with word Count Vectorizer of Unigram and Bigram combination score the best performance on the annotated dataset that is being used.

Download article

email to a friend

Cite This Article

ISSN: 2349-6002
Volume: 9
Issue: 5
PageNo: 294-302

Semi-Supervised Classification of Climate Change & Population Related Tweets based on Hash Tags and Accounts annotation

Available:https://ijirt.org/Article?manuscript=156901

Impact Factor
8.01 (Year 2024)

UGC Approved
Journal no 47859

Join Our IPN

IJIRT Partner Network

Submit your research paper and those of your network (friends, colleagues, or peers) through your IPN account, and receive 800 INR for each paper that gets published.

Join Now