Please use this identifier to cite or link to this item: http://ir.lib.seu.ac.lk/handle/123456789/6168
Full metadata record
DC FieldValueLanguage
dc.contributor.authorShafana, A. R. F.-
dc.contributor.authorNihla, M. I. F.-
dc.contributor.authorMusfira, A. F.-
dc.contributor.authorNaja, M. M. F.-
dc.date.accessioned2022-07-06T10:25:56Z-
dc.date.available2022-07-06T10:25:56Z-
dc.date.issued2022-05-25-
dc.identifier.citationBook of Abstracts - Proceedings of the 10th International Symposium 2022 on "Multidisciplinary Research for Encountering Contemporary Challenges”. 25th May 2022. South Eastern University of Sri Lanka, Oluvil, Sri Lanka. pp. 42.en_US
dc.identifier.isbn978-624-5736-37-9-
dc.identifier.urihttp://ir.lib.seu.ac.lk/handle/123456789/6168-
dc.description.abstractThe proliferation of social media enables the public to express their views and perceptions readily online. Twitter is one such platform that helps in obtaining a huge amount of textual data and performing useful analysis. Sentiment Classification is one such analysis undertaken to gain insights into public opinion on a certain topic. Although this has been prevalently done using many approaches, the limitations still exist in non-English languages. This study aims to compare the use of the lexical-based approach and machine learning-based approach for classifying the Tamil tweets based on their sentiment. Twitter API was used to perform twitter scraping that resulted in 45852 tweets in total. 300 random tweets were then classified to their respective sentiments by subject experts in the field, this annotated data was used as ground truth and 06 underlying studies were performed on the processed and cleaned data. Four machine learning algorithms (Support Vector Machine, eXtreme Gradient Boosting, Random Forest, and Gaussian Naïve Bayes) and two lexical-based analyzers (VADER and TextBlob) were used for this comparative analysis. The results suggested that the machine learning algorithms performed extremely well where the Support Vector Machine secured the best performance score of all. This study serves as empirical evidence for those interested in performing sentiment analysis on Tamil language tweets.en_US
dc.language.isoen_USen_US
dc.publisherSouth Eastern University of Sri Lanka, Oluvil, Sri Lanka.en_US
dc.subjectMachine Learningen_US
dc.subjectSentiment Analysisen_US
dc.subjectLexiconen_US
dc.subjectTwitteren_US
dc.subjectSupervised Learningen_US
dc.titleComparative analysis of machine learning algorithms along with lexical analyzers for sentiment analysis in Tamil Languageen_US
dc.typeArticleen_US
Appears in Collections:10th International Symposium - 2022

Files in This Item:
File Description SizeFormat 
IntSym2022BookofAbstracts-42.pdf348.02 kBAdobe PDFThumbnail
View/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.