INDOBERTWEET DENGAN TEMPORAL ATTENTION MECHANISM UNTUK DETEKSI ISU BENCANA DINAMIS MULTIPLATFORM
DOI:
https://doi.org/10.54314/qtam3c42Kata Kunci:
Disaster Issue Detection; Social Media Text Analysis; Multiplatform Big Data; IndoBERTweet; Temporal Attention.Abstrak
Abstract: The development of disaster-related information in the digital space is very rapid and is often detected earlier through social media than through official government channels. This condition highlights the need for a system capable of detecting disaster-related issues quickly and dynamically across various digital platforms. This study aims to develop a dynamic disaster issue detection model based on IndoBERTweet with a Temporal Attention Mechanism using multiplatform big text data in Indonesia. The research methodology includes collecting textual data from various digital platforms such as Twitter, YouTube, TikTok, Quora, and Medium. The data are processed through preprocessing stages using Natural Language Processing (NLP) techniques, timestamp extraction to obtain temporal information, and the application of fine-grained labeling for more detailed classification of disaster-related issues. Subsequently, the IndoBERTweet model is trained with a Temporal Attention Mechanism to capture the relationship between textual context and temporal dynamics in the development of disaster-related issues. The expected results of this research are a model capable of dynamically detecting disaster-related issues by considering informal language contexts and temporal changes. This model is expected to support early warning systems and data-driven disaster management decision-making in Indonesia.
Keywords: Disaster Issue Detection; Social Media Text Analysis; Multiplatform Big Data; IndoBERTweet; Temporal Attention.
Abstrak: Perkembangan informasi kebencanaan di ruang digital berlangsung sangat cepat dan sering kali lebih dahulu terdeteksi melalui media sosial dibandingkan melalui kanal resmi pemerintah. Kondisi ini menunjukkan perlunya sistem yang mampu mendeteksi isu bencana secara cepat dan dinamis dari berbagai platform digital. Penelitian ini bertujuan mengembangkan model deteksi isu bencana dinamis berbasis IndoBERTweet dengan Temporal Attention Mechanism pada big data teks multiplatform di Indonesia. Metode penelitian meliputi pengumpulan data teks dari berbagai platform digital seperti Twitter, YouTube, TikTok, Quora, dan Medium. Data diproses melalui tahapan preprocessing menggunakan teknik Natural Language Processing (NLP), ekstraksi timestamp untuk memperoleh informasi temporal, serta penerapan fine-grained labeling untuk klasifikasi isu bencana yang lebih rinci. Selanjutnya, model IndoBERTweet dilatih dengan Temporal Attention Mechanism untuk menangkap hubungan antara konteks teks dan dinamika waktu dalam perkembangan isu bencana. Hasil penelitian diharapkan menghasilkan model yang mampu mendeteksi isu bencana secara dinamis dengan mempertimbangkan konteks bahasa informal dan perubahan waktu. Model ini diharapkan mendukung sistem peringatan dini dan pengambilan kebijakan kebencanaan berbasis data di Indonesia.
Kata kunci: Deteksi Isu Bencana; Analisis Teks Media Sosial; Big Data Multiplatform; IndoBERTweet; Temporal Attention.
Unduhan
Referensi
C. Zhang, C. Fan, W. Yao, X. Hu, and A. Mostafavi, “Social media for intelligent public information and warning in disasters: An interdisciplinary review,” Int. J. Inf. Manage., vol. 49, no. 480, pp. 190–207, 2019, doi: 10.1016/j.ijinfomgt.2019.04.004.
X. X. Zhu et al., “Geo-Information Harvesting from Social Media Data,” no. Section II, pp. 1–26, 2022, [Online]. Available: http://arxiv.org/abs/2211.00543
C. Steinmetz et al., “Liking, Tweeting and Posting: An Analysis of Community Engagement through Social Media Platforms,” Urban Policy Res., pp. 1–21, 2020, doi: 10.1080/08111146.2020.1792283.
J. Devlin, M. W. Chang, K. Lee, and K. Toutanova, “BERT: Pre-training of deep bidirectional transformers for language understanding,” NAACL HLT 2019 - 2019 Conf. North Am. Chapter Assoc. Comput. Linguist. Hum. Lang. Technol. - Proc. Conf., vol. 1, pp. 4171–4186, 2019.
Y. Liu et al., “RoBERTa: A Robustly Optimized BERT Pretraining Approach,” no. 1, 2019, [Online]. Available: http://arxiv.org/abs/1907.11692
G. F. Shidik et al., “Indonesian disaster named entity recognition from multi source information using bidirectional LSTM (BiLSTM),” J. Open Innov. Technol. Mark. Complex., vol. 10, no. 3, 2024, doi: 10.1016/j.joitmc.2024.100358.
M. Imran, C. Castillo, F. Diaz, and S. Vieweg, Processing social media messages in Mass Emergency: A survey, vol. 47, no. 4. 2016. doi: 10.1145/2771588.
P. C. Theocharopoulos, P. Anagnostou, S. V. Georgakopoulos, S. K. Tasoulis, and V. P. Plagianakos, “Large language models for efficient topic modeling,” Neural Comput. Appl., vol. 37, no. 29, pp. 24421–24439, 2025, doi: 10.1007/s00521-025-11593-9.
A. Mehmood, M. T. Zamir, M. A. Ayub, N. Ahmad, and K. Ahmad, “A Named Entity Recognition and Topic Modeling-based Solution for Locating and Better Assessment of Natural Disasters in Social Media,” 2024, [Online]. Available: http://arxiv.org/abs/2405.00903
“I NDO BERT WEET : A Pretrained Language Model for Indonesian Twitter with Effective Domain-Specific Vocabulary Initialization,” pp. 10660–10668, 2021.
F. Alam, H. Sajjad, M. Imran, and F. Ofli, “CrisisBench: Benchmarking Crisis-related Social Media Datasets for Humanitarian Information Processing,” Proc. Int. AAAI Conf. Web Soc. Media, vol. 15, pp. 923–932, 2021, doi: 10.1609/icwsm.v15i1.18115.
J. Wang, A. Jatowt, and Y. Cai, “Towards Effective Time-Aware Language Representation: Exploring Enhanced Temporal Understanding in Language Models,” ACM Trans. Web, vol. 19, no. 3, pp. 0–34, 2025, doi: 10.1145/3723352.
C. Reuter, A. L. Hughes, and M. A. Kaufhold, “Social Media in Crisis Management: An Evaluation and Analysis of Crisis Informatics Research,” Int. J. Hum. Comput. Interact., vol. 34, no. 4, pp. 280–294, 2018, doi: 10.1080/10447318.2018.1427832.
“E. Arkhangelskaya, S. Nikolenko DEEP LEARNING FOR NATURAL LANGUAGE PROCESSING: A SURVEY,” 2021.
S. Minaee, N. Kalchbrenner, E. Cambria, N. Nikzad, M. Chenaghlu, and J. Gao, “Deep Learning-Based Text Classification,” ACM Comput. Surv., vol. 54, no. 3, 2022, doi: 10.1145/3439726.
et al., “Sentiment Analysis in Social Media: How Data Science Impacts Public Opinion Knowledge Integrates Natural Language Processing (Nlp) With Artificial Intelligence (Ai),” Am. J. Sch. Res. Innov., vol. 4, no. 1, pp. 63–100, 2025, doi: 10.63125/r3sq6p80.
F. A. Furfari(tony), “The Transformer,” IEEE Ind. Appl. Mag., vol. 8, no. 1, pp. 8–15, 2002, doi: 10.1109/2943.974352.
T. Lin, Y. Wang, X. Liu, and X. Qiu, “A survey of transformers,” AI Open, vol. 3, no. September, pp. 111–132, 2022, doi: 10.1016/j.aiopen.2022.10.001.
Y. Mao et al., “A survey on LoRA of large language models,” Front. Comput. Sci., vol. 19, no. 7, pp. 1–124, 2025, doi: 10.1007/s11704-024-40663-9.
S. Edunov, A. Baevski, and M. Auli, “Pre-trained language model representations for language generation,” NAACL HLT 2019 - 2019 Conf. North Am. Chapter Assoc. Comput. Linguist. Hum. Lang. Technol. - Proc. Conf., vol. 1, pp. 4052–4059, 2019, doi: 10.18653/v1/n19-1409.
K. S. Kalyan, A. Rajasekharan, and S. Sangeetha, “AMMUS : A Survey of Transformer-based Pretrained Models in Natural Language Processing,” pp. 1–42, 2021, [Online]. Available: http://arxiv.org/abs/2108.05542
S. Cahyawijaya et al., “IndoNLG: Benchmark and Resources for Evaluating Indonesian Natural Language Generation,” EMNLP 2021 - 2021 Conf. Empir. Methods Nat. Lang. Process. Proc., pp. 8875–8898, 2021, doi: 10.18653/v1/2021.emnlp-main.699.
G. I. Winata et al., “NusaX: Multilingual Parallel Sentiment Dataset for 10 Indonesian Local Languages,” EACL 2023 - 17th Conf. Eur. Chapter Assoc. Comput. Linguist. Proc. Conf., pp. 815–834, 2023, doi: 10.18653/v1/2023.eacl-main.57.
F. Koto, A. Rahimi, J. H. Lau, and T. Baldwin, “IndoLEM and IndoBERT: A Benchmark Dataset and Pre-trained Language Model for Indonesian NLP,” COLING 2020 - 28th Int. Conf. Comput. Linguist. Proc. Conf., pp. 757–770, 2020, doi: 10.18653/v1/2020.coling-main.66.
A. F. Hidayatullah, R. A. Apong, D. T. C. Lai, and A. Qazi, “Pre-trained language model for code-mixed text in Indonesian, Javanese, and English using transformer,” Soc. Netw. Anal. Min., vol. 15, no. 1, pp. 1–17, 2025, doi: 10.1007/s13278-025-01444-9.
K. Karthikeyan, Z. Wang, S. Mayhew, and D. Roth, “Cross-Lingual Ability of Multilingual Bert: an Empirical Study,” 8th Int. Conf. Learn. Represent. ICLR 2020, pp. 1–12, 2020.
L. Cai, X. Mao, Y. Zhou, Z. Long, C. Wu, and M. Lan, “A Survey on Temporal Knowledge Graph: Representation Learning and Applications,” 2024, [Online]. Available: http://arxiv.org/abs/2403.04782
B. Dhingra, J. R. Cole, J. M. Eisenschlos, D. Gillick, J. Eisenstein, and W. W. Cohen, “Time-Aware Language Models as Temporal Knowledge Bases,” Trans. Assoc. Comput. Linguist., vol. 10, pp. 257–273, 2022, doi: 10.1162/tacl_a_00459.
H. Peng et al., “Streaming Social Event Detection and Evolution Discovery in Heterogeneous Information Networks,” ACM Trans. Knowl. Discov. Data, vol. 15, no. 5, 2021, doi: 10.1145/3447585.
X. Zhou and C. Xu, “Tracing the Spatial-temporal evolution of Events based on Social media data,” ISPRS Int. J. Geo-Information, vol. 6, no. 3, 2017, doi: 10.3390/ijgi6030088.
L. Cai et al., “Structural Temporal Graph Neural Networks for Anomaly Detection in Dynamic Graphs,” Int. Conf. Inf. Knowl. Manag. Proc., no. March, pp. 3747–3756, 2021, doi: 10.1145/3459637.3481955.
T. Sakaki, M. Okazaki, and Y. Matsuo, “Earthquake shakes Twitter users: Real-time event detection by social sensors,” Proc. 19th Int. Conf. World Wide Web, WWW ’10, no. April 2010, pp. 851–860, 2010, doi: 10.1145/1772690.1772777.
S. Vieweg, A. L. Hughes, K. Starbird, and L. Palen, “Microblogging during two natural hazards events: What twitter may contribute to situational awareness,” Conf. Hum. Factors Comput. Syst. - Proc., vol. 2, no. April 2010, pp. 1079–1088, 2010, doi: 10.1145/1753326.1753486.
