Work

Sentiment Analysis of ChatGPT Exploration Based on Opinions on Platform X Using Naïve Bayes Algorithm

Year: 2025

Type: article

Abstract: ChatGPT (chat generative pretrained transformer) developed by OpenAI and launched in 2021, quickly gained widespread attention for its ability to understand and generate human-like text responses. ChatGPT can handle a variety of tasks, including answering questions, solving maths problems, coding, and creating scientific articles or journals. Despite its versatility, concerns about the accuracy of responses to exploratory results of using chatGPT to perform various tasks have arisen, prompting the need for further evaluation. This research uses sentiment analysis to assess public opinion towards ChatGPT, using data from posts on the X app (formerly Twitter), accessed through the X developer API using Naive Bayes classification. Naïve Bayes classification algorithm was applied to categorise the sentiment. Findings showed that of the 3,001 posts analysed, 59.24% expressed positive sentiment, 17.56% negative, and 23.2% neutral. The Naïve Bayes algorithm achieved 79.84% accuracy in this classification task. The results indicate a generally positive public perception of ChatGPT, despite the concerns.

Source: Proceeding International Seminar of Science and Technology

DOI: 10.33830/isst.v4i1.5235

Authors: Denisha Trihapningsari, Armita Widyasuri, Mayang Anglingsari Putri, Ahmad Fatihin

Institutions: Indonesia Open University, National Research and Innovation Agency, Kementerian Ketenagakerjaan Republik Indonesia

Language: English

Topic: Sentiment Analysis and Opinion Mining

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W4409649538

FWCI:

Cites:

Cited by:

Related to: 10