Aspect-Based Sentiment Analysis on Webtoon Reviews Using Ensemble Learning - Research Study

Aug 20, 2026

Upload by Jacob

Rapid growth of online comic platforms like LINE Webtoon generates large volumes of user comments containing diverse opinions about narrative elements. These comments offer rich signals about readers’ perceptions of aspects such as plot, characters, and visuals. This study builds an Aspect-Based Sentiment Analysis framework using ensemble learning to classify sentiment in Indonesian Webtoon reviews. A quantitative experimental pipeline includes data collection, text preprocessing, manual aspect-sentiment annotation, TF-IDF feature extraction, ensemble training with Random Forest and XGBoost, and performance evaluation. Using 1,010 annotated comments across plot, character, and visual aspects, aspect-aware ensembles improve sentiment classification, with the tuned Random Forest achieving 0.6584 accuracy and 0.6541 weighted F1.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Introduction

Background and motivation

Gap in existing studies

Problem formulation and approach

Related work

Overview of ABSA development

Aspect identification and sentiment interpretation

Aspect-Based Sentiment Analysis on Webtoon Reviews Using Ensemble Learning Rival Fakhri Amrullah1,*, Fathoni Mahardika1, Dani Indra Junaedi1 1Department of Informatics, Universitas Sebelas April, Sumedang, Indonesia E-mail: [rivalfakhria@gmail.com](rivalfakhria@gmail.com) Received: November 18, 2025 Accepted for publication: January 6, 2026 Published: May 11, 2026 ABSTRACT The rapid growth of online comic platforms such as LINE Webtoon has produced large volumes of user comments that reflect diverse opinions on storytelling elements. Analyzing these comments provides meaningful insights into reader perceptions of aspects such as plot, characters, and visuals. This study proposes an Aspect-Based Sentiment Analysis (ABSA) framework using ensemble learning to classify sentiment in Indonesian Webtoon reviews. The research follows an experimental quantitative methodology consisting of data collection, text preprocessing, manual annotation of aspects and sentiments, feature extraction using Term Frequency–Inverse Document Frequency (TF-IDF), ensemble model training with Random Forest and XGBoost, and performance evaluation. A total of 1,010 annotated comments were used, covering three aspects, plot, character and visual, and three sentiment categories: negative, neutral, and positive. The results demonstrate that incorporating aspect information enhances sentiment classification performance. While the tuned XGBoost model using TF-IDF features achieved an accuracy of 0.6238 in the text-only scenario, the best performance was obtained by the tuned Random Forest ABSA model, which combined TF-IDF and One-Hot Encoded aspect features and achieved an accuracy of 0.6584 with a weighted F1-score of 0.6541. Class-level analysis shows that neutral comments are the easiest to classify, while negative sentiment remains the most challenging due to informal and context-dependent expressions. A 10-fold cross-validation yielded a mean accuracy of 0.6238 with a standard deviation of 0.0572, indicating stable generalization. These findings highlight the effectiveness of aspect-enhanced ensemble learning for sentiment analysis in Indonesian Webtoon reviews. Keywords: Aspect-Based Sentiment Analysis, Ensemble Learning, Random Forest, XGBoost, Webtoon Reviews I. Introduction The rapid development of digital technology has significantly increased the production of web-based content, including online comic platforms such as LINE Webtoon [1], [2] . Beyond entertainment, Webtoon provides an interactive space where readers express opinions through comment sections [3] . These comments often contain subjective evaluations of narrative elements such as plot, characters, and visuals [4], [5], offering valuable data for sentiment analysis, particularly at the aspect level [6], [7] . However, Indonesian Webtoon comments are characterized by informal language, slang, abbreviations, and code-mixing, which pose substantial challenges for sentiment classification [8],[9] . Most existing sentiment analysis studies focus on document-level or sentence-level classification, overlooking aspect-level distinctions [10], [11], [12] . Such approaches are insufficient for multidimensional reviews like Webtoon comments, where users may simultaneously express different sentiments toward different aspects of a story [13] . Aspect-Based Sentiment Analysis (ABSA) addresses this limitation by assigning sentiment polarity to specific aspects within a text [14], [15] . Despite its effectiveness, ABSA research in Indonesia remains limited, particularly in the digital entertainment domain, with most studies concentrating on e-commerce or application reviews [2],[16] . Moreover, many existing works rely on single classifiers such as Naïve Bayes or Support Vector Machine (SVM), which often struggle with highly variable and informal text [17],[18] . Given the short, sparse, and noisy nature of Indonesian Webtoon comments, selecting an appropriate feature repr

FAQs

What is the main goal of the proposed ABSA framework in this study?

To classify sentiment for Indonesian Webtoon reviews at the aspect level (plot, character, and visual) by using aspect-conditioned sentiment prediction.

How does the study model sentiment and aspects?

It assigns predefined aspect labels as contextual guidance, extracts TF-IDF features from text, and trains ensemble classifiers including Random Forest and XGBoost.

Which model performed best, and what were the key results?

The tuned Random Forest with combined TF-IDF and One-Hot encoded aspect features achieved the best performance: 0.6584 accuracy and 0.6541 weighted F1-score.

Aug 20, 2026

Upload by Jacob