Sentiment Analysis on Food Delivery Services in Malaysia

Food delivery services in Malaysia have become increasingly necessary, especially compared to 10 years ago. Since the outbreak of COVID-19 in 2020, contactless services have been crucial in maintaining social distancing to prevent the virus's spread. Among the various services, Foodpanda, Grab...

Full description

Bibliographic Details
Published in:2024 5th International Conference on Artificial Intelligence and Data Sciences, AiDAS 2024 - Proceedings
Main Author: Azhar A.; Mohamad M.; Zulkifli Z.; Zainuddin N.; Mohtar I.A.
Format: Conference paper
Language:English
Published: Institute of Electrical and Electronics Engineers Inc. 2024
Online Access:https://www.scopus.com/inward/record.uri?eid=2-s2.0-85209639368&doi=10.1109%2fAiDAS63860.2024.10730519&partnerID=40&md5=ee6f50dce0d07a35d43d69c96fcf51d8
Description
Summary:Food delivery services in Malaysia have become increasingly necessary, especially compared to 10 years ago. Since the outbreak of COVID-19 in 2020, contactless services have been crucial in maintaining social distancing to prevent the virus's spread. Among the various services, Foodpanda, Grab Food, and Shopee Food have become particularly popular among Malaysian citizens. The primary objectives of this work are to analyze the emotional responses of Malaysian citizens who use these food delivery services through sentiment analysis and to develop a web-based application that visualizes these emotions. This study utilized numerous tweets from the social media platform Twitter, where users express their personal opinions and thoughts publicly. The tweets were pre-processed and cleaned using Python, and the data visualizations were created using Power BI. Three datasets representing the three main services were analyzed using a support vector machine model. After training and testing the model with the cleaned data, an overall average accuracy of 80.7% was achieved, demonstrating the model's effectiveness in predicting tweet sentiment. However, the study faced limitations, such as the model's inability to detect sarcasm and the lack of abbreviation handling in its dictionary. Future work should expand the data sources to include multiple social media platforms, such as Facebook and Instagram, to gain a broader understanding and achieve more accurate sentiment analysis of the public's opinions. © 2024 IEEE.
ISSN:
DOI:10.1109/AiDAS63860.2024.10730519