Sentiment Analysis towards Full Movie Dirty Vote 2024 in X Using Support Vector Machine Method

Azhari, Fajar (2024) Sentiment Analysis towards Full Movie Dirty Vote 2024 in X Using Support Vector Machine Method. Journal La Multiapp, 5 (4).

Text 1459-Article_Text-11375-2-10-20240826.pdf Download (850kB)

Abstract

Intending to provide a sentiment of the general public towards the X’s “Dirty Vote 2024” movie, this study inspects 1500 tweets scraped from ‘Tweepy’ library using Support Vector Machine (SVM) method. The tweets were preprocessed by text mining process by tokenization, removal of stopwords and word weighting to categorize the tweets into positive and negative sentiments. While using the SVM model, the recognition rate was 86% which means that the model can successfully recognize sentiment patterns in the dataset. However, the model had a 14% overall misclassification rate especially when it came to assessing subtlety or ambiguity in expressions which indicate its weakness in handling complexities inherent in sentiment. Thus, the study affirms a high level of precision in SVM for sentiment analysis The study further pointed out the need to advance advanced natural language processing NLP approaches to enhance the accuracy of models particularly in different real world settings where language adaption is highly volatile. The study also has a relevance to filmmakers and marketers; given that it offered a better understanding of the public response that can help in framing future content creation and advertisement advertisements. Thus, for the increase of accuracy and simply to make the methods more resilient, the future studies should investigate the opportunities of using context-aware embeddings and a hybrid neural network model in the environment of social platforms.

Jenis Item:	Artikel
Subjects:	000 Generalities > 004 Data processing Computer science
Divisions:	Fakultas Sains dan Teknologi > Ilmu Komputer
Pengguna yang mendeposit:	Mrs Siti Masitah
Date Deposited:	06 Feb 2025 08:10
Last Modified:	06 Feb 2025 08:10
URI:	http://repository.uinsu.ac.id/id/eprint/24893

Actions (login required)

View Item