Pengenalan Emosi Ucapan Menggunakan Bidirectional Long Short Term Memory Pada Podcast Malaka
DOI:
https://doi.org/10.30865/json.v7i3.9432Keywords:
Speech Emotion Rcognition , LSTM , BiLSTM, Emosi, PodcastAbstract
Penelitian ini mengembangkan sistem pengenalan emosi ujaran berbahasa Indonesia menggunakan metode Bidirectional Long Short Term Memory (BiLSTM) dengan sumber data berupa ujaran alami dari podcast Malaka. Dataset diperoleh dari audio YouTube yang di konversi ke format WAV, disegmentasi, serta dilabeli emosi marah, netral, sedih dan senang berdasarkan acuan dataset CREMA-D. Tahapan prapemrosesan meliputi silce removal, normalisasi sinyal, dan augmentasi data. Ektrasi ciri dilakukan menggunakan Mel-Frequency Cepstral Coefficients (MFCC) beserta fitur delta dan delta-delta untuk mempresentasikan karakteristik spektral dan temporal sinyal suara. Model Bidirectional Long Short Term Memory dilatih dengan pembagian data 80% data latih, 10% data validasi, dan 10% data uji, serta dioptimasi dengan menggunakan algoritma adam. Hasil pengujian menunjukan bahwa model mencapai akurasi terbaik sebesar 80.32% dengan nilai pecision, recall, dan f1-score yang relatif seimbang pada seluruh kelas emosi. Hasil ini menunjukan bahwa Bidirectional Long Short Term Memory (BiLSTM) efektif dalam memodelkan dinamika temporal emosi pada ujaran podcast berbahasa indonesia.
References
J. Nesi, E. H. Telzer, and M. J. Prinstein, Handbook of Adolescent Digital Media Use and Mental Health. 2022. doi: 10.1017/9781108976237.
F. KASYIDI, R. ILYAS, and N. M. ANNISA, “Peningkatan Kemampuan Pengenalan Emosi Melalui Suara dalam Bahasa Indonesia,” MIND J., vol. 6, no. 2, pp. 194–204, Dec. 2021, doi: 10.26760/mindjournal.v6i2.194-204.
T. B. Putri, S. Saidah, B. Hidayat, F. Qothrunnada, T. Telekomunikasi, and U. Telkom, “Deteksi Emosi Berdasarkan Sinyal Suara Manusia Menggunakan Discrete Wavelet Transform ( DWT ) Dengan Klasifikasi Support Vector Machine ( SVM ) antar manusia tidak selalu terjadi dengan baik , ada beberapa faktor dari interaksi yang dapat berekspresi atau,” vol. 3, no. 1, pp. 1–10, 2023.
S. Helmiyah, A. Fadlil, A. Yudhana, A. Dahlan, and P. Studi Teknik Elektro, “Pengenalan Pola Emosi Manusia Berdasarkan Ucapan Menggunakan Ekstraksi Fitur Mel-Frequency Cepstral Coefficients (MFCC) Speech Based Emotion Pattern Recognition Using Mel-Frequency Cepstral Coefficients (MFCC) Feature Extraction,” Cogito Smart J., vol. 4, no. 2, p. 372, 2018.
F. J. Tanudjaja, E. Y. Puspaningrum, and Y. V. Via, “Klasifikasi Jenis Emosi Melalui Ucapan Menggunakan Metode Convolutional Neural Network,” Teknologi, vol. 13, no. 2, pp. 1–11, 2023, doi: 10.26594/teknologi.v13i2.3740.
H. K. Bhuyan, B. Brahma, N. K. Kamila, S. Peram, B. Leelambika, and A. Sahu, “Mel-Spectrograms Based LSTM Model for Speech Emotion Recognition,” Trait. du Signal, vol. 42, no. 3, pp. 1353–1365, 2025, doi: 10.18280/ts.420312.
R. A. Nawasta, N. H. Cahyana, and H. Heriyanto, “Implementation of Mel-Frequency Cepstral Coefficient as Feature Extraction using K-Nearest Neighbor for Emotion Detection Based on Voice Intonation,” Telematika, vol. 20, no. 1, p. 51, 2023, doi: 10.31315/telematika.v20i1.9518.
N. Aini Lailla Asri, R. Ibnu Adam, and B. Arif Dermawan, “Speech Recognition Untuk Klasifikasi Pengucapan Nama Hewan Dalam Bahasa Sunda Menggunakan Metode Long-Short Term Memory,” JATI (Jurnal Mhs. Tek. Inform., vol. 7, no. 2, pp. 1242–1247, 2023, doi: 10.36040/jati.v7i2.6744.
N. D. Pah, PEMROSESAN SINYAL DIGITAL, Edisi Pert. Yogyakarta: Graha Ilmu, 2018.
C. Roy et al., “Stacked convolutional neural network for emotion recognition using multi feature speech analysis,” Sci. Rep., vol. 15, no. 1, Dec. 2025, doi: 10.1038/s41598-025-28766-0.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Jurnal Sistem Komputer dan Informatika (JSON)

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

This work is licensed under a Creative Commons Attribution 4.0 International License
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under Creative Commons Attribution 4.0 International License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (Refer to The Effect of Open Access).

