Salsabila Rahadatul Aisy, Budi Prasetiyo

Sentiment Analysist of the TPKS Law on Twitter Using InSet Lexicon with Multinomial Naïve Bayes and Support Vector Machine Based on Soft Voting

Introduction

Sentiment analysist of the tpks law on twitter using inset lexicon with multinomial naïve bayes and support vector machine based on soft voting. Analyze Twitter sentiment on Indonesia's TPKS Law using InSet Lexicon, Multinomial Naïve Bayes, & Support Vector Machine with Soft Voting, achieving 84.31% accuracy.

Abstract

Abstract. The Indonesian Sexual Violence Law (TPKS Law) is a law that regulates forms of sexual violence. The TPKS Law reaped pros and cons in the drafting process and was officially ratified on April 12th, 2022. However, after being ratified, pros and cons can still be found and supervision is needed over the implementation of the law. Purpose: This study was conducted to identify the application and accuracy of soft voting on multinomial naïve Bayes and support vector machine algorithm, also to find out public opinion on the TPKS Law as a support tool in evaluating the law. Methods/Study design/approach: The method used is InSet lexicon for labeling with the soft voting classification method on the multinomial naive Bayes and support vector machine algorithm. Result/Findings: The accuracy obtained by applying 10 k-fold cross validation in soft voting is 84.31%, which uses a weight of 1:3 for multinomial naive Bayes and support vector machines. Soft voting obtains better accuracy than its standalone predictor, and also works well for sentiment analysis of the TPKS Law. Novelty/Originality/Value: This study using two combined lexicons (Colloquial Indonesian lexicon and the InaNLP formalization dictionary) in normalization process and using InSet lexicon as automatic labeling for sentiment analysis on TPKS Law.

Review

This paper presents a sentiment analysis of the Indonesian Sexual Violence Law (TPKS Law) using Twitter data, a highly relevant and socially impactful topic in Indonesia. The authors aim to evaluate the application and accuracy of a soft voting ensemble combining Multinomial Naïve Bayes (MNB) and Support Vector Machine (SVM) algorithms, while also seeking to gauge public opinion on the law to support its evaluation. This study addresses a pertinent real-world problem, employing established computational methods to derive insights from social media, which is a valuable approach for understanding public discourse around significant legislative actions. A key strength of the study lies in its methodological approach, particularly the use of a soft voting ensemble for sentiment classification. The reported accuracy of 84.31% using 10-fold cross-validation with a 1:3 weight for MNB and SVM suggests a robust performance, and the claim that soft voting outperforms standalone predictors is a promising result for the chosen technique. The novelty aspect, emphasizing the combination of Colloquial Indonesian and InaNLP lexicons for normalization, along with the application of the InSet lexicon for automatic labeling specifically on TPKS Law sentiment, highlights a targeted and domain-specific approach to addressing the nuances of Indonesian language in this context. Despite these strengths, the abstract leaves several critical questions unanswered, particularly regarding the substantive findings. While the purpose includes "to find out public opinion on the TPKS Law," the results section primarily focuses on the *method's accuracy* and *performance*, without offering any concrete insights into what that public opinion actually entails (e.g., the distribution of positive, negative, or neutral sentiments). Furthermore, crucial details about the data, such as its size, collection period, or specific characteristics of the Twitter corpus, are absent, which impedes a full understanding of the study's scope and generalizability. Elaboration on the rationale behind the 1:3 weighting for the soft voting ensemble and a direct comparison of the individual MNB and SVM accuracies against the ensemble's performance would also strengthen the methodological claims.

Full Text

Comments

You need to be logged in to post a comment.