Klasifikasi Dukungan Politik Berdasarkan Analisis Heatmap Aktivitas Komentar di Facebook Menggunakan CNN 1D
Main Article Content
Wahyudin Ahmadi
Bramasto Daru Satrio
Suparno
Analisis sentimen opini publik di media sosial menghadapi tantangan keterbatasan data (small dataset) dan ketidakseimbangan kelas (class imbalance) yang dapat menyebabkan overfitting pada model deep learning. Penelitian ini mengkaji performa arsitektur Convolutional Neural Network 1-Dimensi (CNN-1D) untuk klasifikasi sentimen opini publik terhadap kinerja pemerintahan Presiden Prabowo Subianto ke dalam tiga kategori: Netral, Keluhan/Aduan, dan Apresiasi. Dataset terdiri dari 434 komentar media sosial dengan dominasi kategori keluhan. Untuk mengatasi overfitting dan bias kelas, penelitian ini menerapkan reduksi dimensi embedding, regularisasi dropout (0.5), serta pembobotan kelas (class weight) berdasarkan inverse class frequency. Pengujian dilakukan dengan pembagian data 80:20 dan early stopping untuk mencegah overfitting. Hasil evaluasi menunjukkan akurasi 70.11% dengan presisi tertinggi pada kategori keluhan (81.35%), sementara kategori netral masih mengalami kesalahan klasifikasi akibat keterbatasan sampel dan tumpang tindih semantik. Pembahasan menyoroti perbandingan teoretis CNN-1D dengan Support Vector Machine (SVM) pada dataset berskala kecil serta analisis kesalahan prediksi melalui grafik heatmap confusion matrix
Asudani, D. S., Nagwani, N. K., & Singh, P. (2023). Impact of word embedding models on text analytics in deep learning environment: a review. Artificial Intelligence Review, 56(9), 10345. https://doi.org/10.1007/s10462-023-10419-1
Boulesnane, A., Meshoul, S., & Aouissi, K. (2022). Influenza-like Illness Detection from Arabic Facebook Posts Based on Sentiment Analysis and 1D Convolutional Neural Network. Mathematics, 10(21), 4089. https://doi.org/10.3390/math10214089
Chen, Z., Feng, A., & He, J. (2019). Text sentiment classification based on 1D convolutional hybrid neural network. Journal of Computer Applications, 39(7), 1936–1941. https://doi.org/10.11772/j.issn.1001-9081.2018122477
Fahrezy, A. R., Refina, S. A., Chandra, J. G., & Yora, S. (2025). Sentiment Analysis of Indonesian Hashtag #kaburajadulu Using CNN. 2025 International Conference on Information Technology and Computing (ICITCOM), 222–227. https://doi.org/10.1109/ICITCOM64883.2025.11265040
Habib, M. A., Akhand, M. A. H., & Kamal, M. A. S. (2022). Emotion Recognition from Microblog Managing Emoticon with Text and Classifying using 1D CNN. Journal of Computer Science, 18(12), 1170–1178. https://doi.org/10.3844/jcssp.2022.1170.1178
Huana, H., Guoa, Z., Caib, T., & Heb, Z. (2022). A text classification method based on a convolutional and bidirectional long short-term memory model. Connection Science, 34(1), 2108–2124. https://doi.org/10.1080/09540091.2022.2098926
Ilyas, R., Kasyidi, F., & Eriyadi, M. N. (2024). Penanganan Overfitting pada Klasifikasi Berita Hoax berbasis Neural Networks dengan Dropout dan Regularization. Jurnal Teknologi Dan Komputer, 10(2). https://doi.org/10.31294/jtk.v10i2.23121
Kiranyaz, S., Avci, O., Abdeljaber, O., Ince, T., Gabbouj, M., & Inman, D. J. (2021). 1D convolutional neural networks and applications: A survey. Mechanical Systems and Signal Processing, 151, 107398. https://doi.org/10.1016/j.ymssp.2020.107398
Nugraha, W., & al., et. (2025). Public Sentiment Toward the Indonesian Capital Relocation Policy on X Using a BiLSTM-CNN Model. Telematika, 20(2). https://doi.org/10.61769/telematika.v20i2.796
Pang, B., & Lee, L. (2008). Opinion Mining and Sentiment Analysis. Foundations and Trends in Information Retrieval, 2(1–2), 1–135.
Pham, H., & al., et. (2021). AutoDropout: Learning Dropout Patterns to Regularize Deep Networks. AAAI 2021. https://arxiv.org/abs/2101.01761
Riwaldi, M. R. F., & Aripin, A. (2026). Hybrid CNN-BiLSTM untuk Analisis Sentimen Multi-Platform terhadap Insiden Keamanan Pangan Program Makan Bergizi Gratis. BIT (Budi Luhur Information Technology), 8(1). https://doi.org/10.32493/bit.v8i1.9896
Rohalia, A., & Rahmika, A. R. (2026). Leveraging Convolutional Neural Networks and Random Forests for Advanced Sentiment Classification of Social Media Responses on Public Services. JAIC (Journal of Artificial Intelligence and Computing), 10(1). https://doi.org/10.30871/jaic.v10i1.11965
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., & Salakhutdinov, R. (2014). Dropout: A Simple Way to Prevent Neural Networks from Overfitting. Journal of Machine Learning Research, 15(1), 1929–1958.
Talai, Z., & Kherici, N. (2023). Text Classification and Sentiment Analysis using 1D CNN. Information Processing at the Digital Age Journal, 27(2), 46–51.
Tan, Z., Chen, J., Kang, Q., Zhou, M., Abusorrah, A., & Sedraoui, K. (2022). Dynamic Embedding Projection-Gated Convolutional Neural Networks for Text Classification. IEEE Transactions on Neural Networks and Learning Systems, 33(3), 973–982. https://doi.org/10.1109/TNNLS.2020.3036192
Wen, L., & al., et. (2021). Maximum Pooling Dropout with Weight Attenuation for Overfitting in CNN. Wireless Communications & Mobile Computing. https://doi.org/10.1155/2021/8493795
Yilmaz, S. F., Kaynak, E. B., Koç, A., Dibeklioğlu, H., & Kozat, S. S. (2021). Multi-Label Sentiment Analysis on 100 Languages with Dynamic Weighting for Label Imbalance. IEEE Transactions on Neural Networks and Learning Systems, 33(1), 331–343. https://doi.org/10.1109/TNNLS.2021.3094304
