Claim Missing Document
Check
Articles

Found 32 Documents
Search

Implementasi Algoritma Naïve Bayes dan Decision Tree pada Sentimen Dampak Bencana Hidrometoeorologi Arry Wahyu Setiawan; Suprianto Suprianto; Yunianita Rahmawati; Suhendro Busono
Jutisi : Jurnal Ilmiah Teknik Informatika dan Sistem Informasi Vol. 15 No. 4 (2026): Agustus 2026
Publisher : STMIK Banjarbaru

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35889/jutisi.v15i4.3858

Abstract

Abstract Hydrometeorological disasters often trigger public responses on social media, so it’s requiring analysis public opinion toward such disaster.  This study compares the performance of Decision Tree and Naïve Bayes Algorithm to analyze performance on analyzing sentiments of respond related to hydrometeorological disaster. The dataset that used contains of 11890 comments. Comments labeled into Positive, Negative, and neutral then processed by preprocessing, TF-IDF weighting, and classification with 5-fold cross validation. Validation results Naïve Bayes had mean accuracy by 0.79 which higher than Decision Tree, which scored 0,65. Naïve Bayes also got 0.79 score on accuracy, precision, recall, and F1-Score compares to Decision tree that scored scored 0.63, 0.70, 0.63, and 0.64 respectively. Because the gap of the mean accuracy and test accuracy wasn’t far off, concludes that Naïve Bayes have better generalization than Decision Tree. This result shows that Naïve Bayes is better than Decision Tree on classification of Hydrometeorological Disaster respond. Key Words: Sentiment Analysis; Hydrometeorology Disaster; Naïve Bayes; Decision tree; TF-IDF; Social Media   Abstrak Bencana hidrometeorologi memicu berbagai respons masyarakat di media sosial. maka diperlukan analisis untuk memahami opini publik terhadap bencana terkait. Tujuan penelitian ini adalah membandingkan kinerja algoritma Naïve Bayes dan Decision Tree dalam melakukan analisis sentimen terkait bencana Hidrometeorologi. Data digunakan berasal dari TikTok, Instagram Facebook yang terdiri dari 11.890 komentar dengan label kelas positif, negatif, dan netral yang kemudian diproses melalui tahapan pra-pemrosesan, pembobotan TF-IDF, dan klasifikasi dan validasi 5-fold cross validation. Hasil validasi menunjukkan Naïve Bayes memperoleh mean accuracy 0,79 lebih tinggi daripada Decision Tree yang mendapat skor 0,65. Pada data uji, Naïve Bayes memperoleh accuracy, precision, recall, dan F1-Score yang seragam 0,79 lebih unggul dari Decision Tree dengan accuracy 0,63, precision 0,70, recall 0,63, dan F1-score 0,64. Dikarenakan jarak mean acmecuracy dan test accuracy berdekatan disimpulkan Naïve Bayes mempunyai kemampuan generalisasi lebih baik daripada Decision Tree. Hasil ini menunjukkan Naïve Bayes unggul dalam mengklasifikasi sentimen terkait bencana Hidrometeorologi
Evaluation of Multimedia Branding Strategies for Postgraduate Management Enrollment and Institutional Image: Evaluasi Strategi Pencitraan Berbasis Multimedia untuk Pendaftaran Program Manajemen Pascasarjana dan Citra Institusi Faris Dwi Prasetya; Suhendro Busono
Procedia of Engineering and Life Science Vol. 10 No. 1 (2026): Proceedings of the 9th Seminar Nasional Sains 2026
Publisher : Universitas Muhammadiyah Sidoarjo

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.21070/pels.v10i1.1884

Abstract

This research aims to analyze the effectiveness of branding strategies in increasing the number of enthusiasts in the Master of Management Study Program at Muhammadiyah University of Sidoarjo (UMSIDA). Using qualitative methods, this research explores the perceptions and experiences of various stakeholders regarding the branding efforts that have been carried out. Data was collected through in-depth interviews with study program leaders, marketing staff, students and alumni, as well as analysis of documents related to branding strategies that have been implemented. The research results show that effective branding strategies include strengthening the image of study programs through social media, collaborating with industry, and improving the quality of academic services. However, several challenges were also found, such as limited resources and intense competition in the higher education market. This research provides recommendations for improving branding strategies to increase the attractiveness of the UMSIDA Master of Management Study Program in the future.