BIG DATA APPROACH TO SENTIMENT ANALYSIS IN MACHINE LEARNING-BASED MICROBLOGS - PERSPECTIVES OF RELIGIOUS MODERATION PUBLIC POLICY IN INDONESIA

Aug 6, 2026

Religious moderation is framed through moderate thinking and understanding, moderate behavior, and moderate religious worship. Leveraging information technology, the study evaluates Indonesia’s public policies on religious moderation reflected in microblog discussions initiated by the Ministry of Religion, including changes to the halal logo, shifts in halal certification authority, and mosque loudspeaker volume regulations. Big-data sentiment extraction is performed on Twitter, Instagram, and YouTube using six machine learning classifiers, with Gradient Boosting achieving the highest accuracy of 82.27%.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Abstract

1. Introduction

2. Literature Review

BIG DATA APPROACH TO SENTIMENT ANALYSIS IN MACHINE LEARNING-BASED MICROBLOGS: PERSPECTIVES OF RELIGIOUS MODERATION PUBLIC POLICY IN INDONESIA Mhd. Furqan1*, Ahmad Fakhri Ab. Nasir2 Faculty of Science and Technology, Universitas Islam Negeri Sumatera Utara, Indonesia 1 Faculty of Computing, University Malaysia Pahang Al-Sultan Abdullah, Pekan, Malaysia2 [mhdfurqan@uinsu.ac.id](mhdfurqan@uinsu.ac.id1)[1](mhdfurqan@uinsu.ac.id1), [afakhri@umpsa.edu.my](afakhri@umpsa.edu.my2)[2](afakhri@umpsa.edu.my2) Received : 21 February 2024, Revised: 03 May 2024, Accepted : 05 May 2024 *Corresponding Author ABSTRACT The concept of religious moderation encompasses three key aspects, namely moderate thinking and understanding, moderate behavior, and moderate religious worship. With advancements in information technology, people now have the means to express their opinions through microblogs, pertaining to issues of religious moderation initiated by the Ministry of Religion of Indonesia. This study aims to evaluate public policies introduced by the Ministry of Religion regarding religious moderation such as changes in the halal logo, transfer of authority for halal certification, and regulations on the volume of loudspeakers in the mosque. Public opinions collected as the big data to get the information about public sentiment with those issues. Sentiment analysis was conducted on three primary microblogs such as Twitter, Instagram and YouTube using six machine learning algorithms. These include Naïve Bayes, Support Vector Machine (SVM), k-Nearest Neighbor (k-NN), Bagging Classifier, Random Forest, and Gradient Boosting Classifier. The test results showed the highest accuracy is Gradient Boosting reached 82.27%. Keywords: Religious moderation, sentiment analysis, microblogs, big data, machine learning. 1. Introduction Social media has become an integral part of daily life due to its close association with various activities. These activities include searching for product reviews, sharing information, seeking public opinion, etc. For example, business executives consider social media an indispensable marketing and advertising tool (Asif, 2020) . Government decision-makers rely on social media to gauge public sentiment, which can influence future public policies (Barghout, 2015) . Social media encompasses various forms, such as microblogs that facilitate easy information sharing through short sentences and digital photos. Additionally, website links allow users to share content like articles or news about current events effortlessly. Videos able also be embedded on websites or blogs, creating an illusion which they are hosted by the author of the respective platform. The development of big data has provided vast opportunities to sentiment analysis, particularly in microblogging data (Khan & Malviya, 2021) . Sentiment analysis involves contextually mining text to identify, extract, and analyze subjective information related for various entities, such as topics, products, and services(Pang & Lee, 2008) . The rapid growth of internet-based applications, including web pages, social networks, and blogs, has empowered people to express opinions and comments regarding products, services, and daily activities. Sentiment analysis is a powerful tool to businesses and governments for gauge public opinion on products or services, aiding decision-making processes (Setia & Rahman, 2022) . Numerous opinions and sentiments flood social media platforms every day, prompting the have to a data driven approach for conducting sentiment analysis. This approach, known as microblogging, encompasses various ways in which people express their opinions through textbased comments (Joachims, 1998) . The classifier in supervised machine learning that makes use of data that has been manually annotated or that has been labeled with a noisy indication, like emoticons (Maity, 2016) . In this context, Support Vector Machines (SVM), Naïve Bayes, and k-Nearest Neig

FAQs

What dimensions of religious moderation does the study focus on?

It covers moderate thinking and understanding, moderate behavior, and moderate religious worship.

Which Indonesian religious moderation public policy issues are analyzed in microblogs?

The study examines halal logo changes, transfer of authority for halal certification, and regulations on mosque loudspeaker volume.

Which machine learning algorithm performed best in the sentiment analysis?

Gradient Boosting produced the highest accuracy, reaching 82.27%.

Aug 6, 2026

Upload by Olivia Brown