Article Per Year (5 Year)
p-Index From 2021 - 2026
0.444
P-Index
This Author published in this journals
All Journal Journal of Information Systems and Informatics Jurnal Komputer, Informasi dan Teknologi
M. Fajar Ramadhan
Satu University
Computer Science & IT Education Languange, Linguistic, Communication & Media

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Analisis Sentimen Kutipan Media Sosial Berbahasa Indonesia Menggunakan Convolutional Neural Network M. Fajar Ramadhan; Febriyanti Panjaitan; Winarnie; Hery Oktafiandi; Yohanes
Jurnal Komputer, Informasi dan Teknologi Vol. 5 No. 2 (2025): Desember
Publisher : Penerbit Jurnal Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.53697/jkomitek.v6i1.3627

Abstract

Media sosial menjadi salah satu ruang utama ekspresi opini dan emosi masyarakat Indonesia. Berbagai pandangan, keluhan, dukungan, maupun refleksi pribadi sering dituangkan dalam bentuk teks pendek seperti caption dan kutipan (quotes). Teks-teks ini tidak hanya memuat informasi, tetapi juga sentimen yang dapat menggambarkan sikap pengguna terhadap suatu isu. Penelitian ini bertujuan untuk mengembangkan model analisis sentimen kutipan media sosial berbahasa Indonesia menggunakan Convolutional Neural Network (CNN) dan membandingkannya dengan pendekatan klasifikasi klasik berbasis Term Frequency–Inverse Document Frequency (TF-IDF) dan Logistic Regression. Data yang digunakan berasal dari dataset terbuka opini publik di media sosial terkait pasar saham Indonesia. Dari dataset tersebut dipilih secara seimbang 1.200 kutipan berbahasa Indonesia, masing-masing 400 berlabel negatif, 400 netral, dan 400 positif. Tahapan penelitian meliputi pembersihan teks, normalisasi, tokenisasi, pembagian data latih dan uji (80:20), pembentukan vektor TF-IDF sebagai baseline, pembentukan urutan indeks kata untuk CNN, perancangan arsitektur text-CNN dengan tiga kernel konvolusi paralel, serta evaluasi model menggunakan akurasi, precision, recall, F1-score, dan confusion matrix. Model baseline TF-IDF + Logistic Regression menghasilkan akurasi 66% dengan macro F1-score 0,66 pada data uji. CNN varian awal (CNN V1) memberikan performa sebanding dengan akurasi 65% dan macro F1-score 0,65, namun menunjukkan gejala overfitting. Setelah kapasitas model dikurangi dan regularisasi diperkuat (CNN V2), kinerja meningkat menjadi akurasi 67,5% dan macro F1-score 0,67, sedikit melampaui baseline klasik. Hasil ini mengindikasikan bahwa CNN mampu menangkap pola n-gram lokal pada kutipan secara efektif, meskipun keunggulannya terhadap pendekatan klasik masih terbatas pada skala data yang relatif kecil.
Machine Learning Classification of SCD, CHF, and NSR Using 15-Minute ECG-Derived HRV Features Febriyanti Panjaitan; Win Ce; M. Fajar Ramadhan; Winarnie; Hery Oktafiandi
Journal of Information System and Informatics Vol 8 No 2 (2026): April
Publisher : Asosiasi Doktor Sistem Informasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63158/journalisi.v8i2.1557

Abstract

Heart disease remains one of the leading causes of mortality worldwide, making early detection essential for effective intervention. Heart Rate Variability (HRV) is widely used as a non-invasive marker for assessing cardiac conditions, and machine learning has shown potential in classifying heart diseases such as Sudden Cardiac Death (SCD) and Congestive Heart Failure (CHF). This study evaluates the performance of Support Vector Machine (SVM), Decision Tree (DT), and K-Nearest Neighbors (KNN) using 15-minute ECG signals comprising three 5-minute segments. The dataset consists of 53 subjects, generating 159 segments, including SCD, CHF, and Normal Sinus Rhythm (NSR). To prevent data leakage, a subject-wise split (80:20) is applied for training and testing. Two evaluation scenarios are considered: per-segment classification and combined 15-minute classification. Results indicate that SVM and DT achieve consistently high, stable performance with near-perfect accuracy, precision, recall, and F1-score, whereas KNN shows lower, more variable performance, particularly in segment-based analysis. The combined 15-minute approach provides more stable results, suggesting improved HRV representation and class separability. Although the results are promising, further validation with larger, more diverse datasets is required to ensure robustness and generalizability. This study highlights the potential of HRV-based machine learning while emphasizing the importance of appropriate temporal representation and rigorous evaluation design.
Co-Authors Febriyanti Panjaitan Hery Oktafiandi Win Ce Winarnie Yohanes