Indonesian Journal of Informatic Research and Software Engineering (IJIRSE) https://journal.irpi.or.id/index.php/ijirse <p>Indonesian Journal of Informatic Research and Software Engineering (IJIRSE) is a scientific journal published by the Indonesian Research and Publication Institute (IRPI) in collaboration with several universities throughout Riau and Indonesia. IJIRSE will be published 2 (two) times a year, March and september, each edition containing 10 (ten) articles. Articles may be written in Indonesian or English. articles are original research results with a maximum plagiarism of 20%. Articles submitted to IJIRSE will be reviewed by at least 2 (two) reviewers. The entire process until IJIRSE is published will be free of charge. IJIRSE is registered in National Library with Number International Standard Serial Number (ISSN) Online <a href="http://issn.pdii.lipi.go.id/issn.cgi?daftar&amp;1613710054&amp;1&amp;&amp;" target="_blank" rel="noopener">2775-5754</a> and Print <a href="https://issn.lipi.go.id/terbit/detail/20210528340787323">2797-2712</a>.</p> Institut Riset dan Publikasi Indonesia en-US Indonesian Journal of Informatic Research and Software Engineering (IJIRSE) 2797-2712 Information Gain Feature Selection for Temporal Sentiment Analysis of Pedulilindungi Application Review using Naïve Bayes Classifier Algorithm https://journal.irpi.or.id/index.php/ijirse/article/view/2217 <p><em>Through the Instruction of the Minister of Home Affairs of the Republic of Indonesia Number 38 of 2021 concerning the Implementation of Restrictions on Community Activities (PPKM), all communities are required to use the Pedulilindungi application from August 31, 2021, to September 6, 2021, and updated regularly. Users can download and access the Pedulilindungi application through the Google Play Store application market. There, users can directly assess an application by providing reviews that can describe user responses and satisfaction with the application. The Naïve Bayes Classifier (NBC) algorithm is applied to perform modeling in classifying temporal sentiment analysis data. Prior to classification, a feature selection process with information gain is performed. Based on the experimental results, the best evaluation was produced on temporal data dated September 03, 2021, with an accuracy of 91.9% and precision and recall values of 99.9% and 91.9%, respectively.</em></p> Siti Syahidatul Helma Dini Hidayatul Qudsi Ivan Chatisa Copyright (c) 2025 Indonesian Journal of Informatic Research and Software Engineering (IJIRSE) 2025-08-05 2025-08-05 5 2 98 106 10.57152/ijirse.v5i2.2217