Explainable Industry-Aware Social Media Virality Prediction Using Temporal and Categorical Features

Muhammad Ariqoh Bahrul Ulum; Riska Dhenabayu; Naufal Rifqi  Nasution; Mohamad Syamsul Hidayat; Hujjatullah Fazlurrahman

doi:10.59395/ijadis.v7i3.1647

Authors

Muhammad Ariqoh Bahrul Ulum Universitas Negeri Surabaya
Riska Dhenabayu Universitas Negeri Surabaya https://orcid.org/0000-0002-6530-2099
Naufal Rifqi Nasution Universiti Utara Malaysia
Mohamad Syamsul Hidayat Universitas Islam Majapahit
Hujjatullah Fazlurrahman Universitas Negeri Surabaya

DOI:

https://doi.org/10.59395/ijadis.v7i3.1647

Keywords:

Virality Prediction , Explainable AI , Industry-Aware, Temporal Features , Categorical Features

Abstract

This study proposes an explainable, industry-aware machine learning framework to predict social media post virality using pre-publication metadata. Utilizing 528 observations (January 2025–February 2026) from Instagram and TikTok across the IT, F&B, and manufacturing sectors, virality was defined using a leakage-free 75th-percentile engagement threshold. Several classifiers, Logistic Regression, Random Forest, SVM, XGBoost, and CatBoost, were evaluated using five-fold stratified cross-validation and time-based testing. Model interpretability was addressed using SHAP analysis, and optimal posting times were simulated via a marginal approach.During cross-validation, the categorical-only XGBoost achieved an F1-score of 0.4158 and ROC-AUC of 0.6046, while the operational simulation model (combined_main_no_client) yielded a lower F1-score of 0.3581. In chronological final-test evaluation (January–February 2026), the model suffered severe performance decay due to temporal covariate shift, dropping to an F1-score, Recall, and Precision of 0.2917, while missing 70.83% of actual viral cases. SHAP analysis identified month index and IT industry affiliation as the strongest predictive features. These findings highlight the challenges of pre-publication virality forecasting and emphasize the vulnerability of static classifiers to temporal shifts in low-sample environments.

88 30

Downloads

Download data is not yet available.

References

[1] Yashen Wang; Dongyang Li; Jianshan Sun; Yuan Kun; Yinan Jiang; Yin Zhang; Jie Cao, Combining macro and micro: feature-driven dynamic graphlearning for social media popularity prediction, 2025, doi: https://doi.org/10.1007/s11280-024-01323-x. DOI: https://doi.org/10.1007/s11280-024-01323-x

[2] N. Chaudhuri, G. Gupta, dan A. Popovi, Do you believe it? Examining user engagement with fake news on social media platforms, Technol. Forecast. Soc. Change, vol. 212, hlm. 123950, Mar 2025, doi: 10.1016/j.techfore.2024.123950. DOI: https://doi.org/10.1016/j.techfore.2024.123950

[3] B. Zhang dkk., CasDacGCN: A Dynamic Attention-Calibrated Graph Convolutional Network for Information Popularity Prediction, Entropy, vol. 27, no. 10, hlm. 1064, Okt 2025, doi: 10.3390/e27101064. DOI: https://doi.org/10.3390/e27101064

[4] F. A. Zahrah, R. Dhenabayu, M. F. W. Rahman, R. S. Dewi, dan Z. H. Aminah, Predicting Social Media Post Engagement and Virality Using Graph Neural Network Approaches and Content-Based Features, Kinet. Game Technol. Inf. Syst. Comput. Netw. Comput. Electron. Control, Jun 2026, doi: 10.22219/kinetik.v11i3.2686. DOI: https://doi.org/10.22219/kinetik.v11i3.2686

[5] M. Esteban-Bravo, L. d. l. M. Jimnez-Rubido, dan J. M. Vidal-Sanz, Predicting the virality of fake news at the early stage of dissemination, Expert Syst. Appl., vol. 248, hlm. 123390, Agu 2024, doi: 10.1016/j.eswa.2024.123390. DOI: https://doi.org/10.1016/j.eswa.2024.123390

[6] Riovan Styx Roring*, Chih How Bong, Narayanan Kulathuramaiyer, From Data Mining to Predictive Analytics: Progress in Understanding and Forecasting Social Media Virality | International Research Journal of Multidisciplinary Scope (IRJMS), Jan 2026, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://www.irjms.com/journal/from-data-mining-to-predictive-analytics-progress-in-understanding-and-forecasting-social-media-virality/ DOI: https://doi.org/10.47857/irjms.2026.v07i01.06486

[7] Yanchao Liu, Junpeng Gong, Wenchao Song, Chi Zhang, Pengzhou Zhang, TEGKT: tendency-enhanced evolution graph KAN transformer for information popularity prediction, 2025, doi: https://doi.org/10.1007/s44443-025-00170-8. DOI: https://doi.org/10.1007/s44443-025-00170-8

[8] Lazebnik & Iny, Temporal graphs anomaly emergence detection: benchmarking for social media interactions | Applied Intelligence | Springer Nature Link, 2024, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://link.springer.com/article/10.1007/s10489-024-05821-3#citeas

[9] Yuxuan Du & Jing Lu, The auto-correlation enhanced transformer for information diffusion prediction in social networks | Evolving Systems | Springer Nature Link, 2025, Diakses: 10 Juni 2026. [Daring]. Tersedia pada: https://link.springer.com/article/10.1007/s12530-025-09672-2#citeas

[10] Mingjun Xi, Pingshan Liu, Seshu Yu, Temporal information-aware multimodal learning network for user-generated video popularity prediction, 2025, doi: https://doi.org/10.1007/s00530-025-02042-7. DOI: https://doi.org/10.1007/s00530-025-02042-7

[11] Anggraeni et al., DETECTION OF MICRO-VIRAL CONTENT ON TIKTOK THROUGH SOCIAL LISTENING AND MACHINE LEARNING | JITK (Jurnal Ilmu Pengetahuan dan Teknologi Komputer), 2026, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://ejournal.nusamandiri.ac.id/index.php/jitk/article/view/7472 DOI: https://doi.org/10.33480/jitk.v11i3.7472

[12] C.-Y. Sang, J.-J. Chen, dan S.-G. Liao, DyHGTCR-Cas: Learning unified spatio-temporal features based on dynamic heterogeneous graph neural network for information cascade prediction, Inf. Process. Manag., vol. 62, no. 3, hlm. 104029, Mei 2025, doi: 10.1016/j.ipm.2024.104029. DOI: https://doi.org/10.1016/j.ipm.2024.104029

[13] A. Riazi dan Y. Wang, Differentiating broadcast from viral: a causal inference approach for information diffusion analysis, Appl. Intell., vol. 54, no. 21, hlm. 1037410385, Nov 2024, doi: 10.1007/s10489-024-05723-4. DOI: https://doi.org/10.1007/s10489-024-05723-4

[14] Zhixuan Xu; Minghui Qian, Predicting Popularity of Viral Content in Social Media through a Temporal-Spatial Cascade Convolutional Learning Framework, 2023, doi: https://doi.org/10.3390/math11143059. DOI: https://doi.org/10.3390/math11143059

[15] Kim et al., Frontiers | How to Catch Customers Attention? A Study on the Effectiveness of Brand Social Media Strategies in Digital Customer Engagement, 2023, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://www.frontiersin.org/journals/psychology/articles/10.3389/fpsyg.2021.800766/full#cite

[16] N. Jalli, Viral Justice: TikTok Activism, Misinformation, and the Fight for Social Change in Southeast Asia, Soc. Media Soc., vol. 11, no. 1, hlm. 20563051251318122, Jan 2025, doi: 10.1177/20563051251318122. DOI: https://doi.org/10.1177/20563051251318122

[17] Zhongya Han, Xiangtang Chen, Kepao Miao, Xiaoxiang Wang, Qinlin Li, Predicting product information diffusion for sustainable quality management in industry 4.0: an improved Bass model approach, 2025, doi: https://doi.org/10.1038/s41598-025-09129-1. DOI: https://doi.org/10.1038/s41598-025-09129-1

[18] Sangiorgio et al., Evaluating the effect of viral posts on social media engagement | Scientific Reports, 2025, Diakses: 10 Juni 2026. [Daring]. Tersedia pada: https://www.nature.com/articles/s41598-024-84960-6#citeas

[19] H. He dan E. A. Garcia, Learning from Imbalanced Data, IEEE Trans. Knowl. Data Eng., vol. 21, no. 9, hlm. 12631284, Sep 2009, doi: 10.1109/TKDE.2008.239. DOI: https://doi.org/10.1109/TKDE.2008.239

[20] Ruidong Jin, Xin Liu, Tsuyoshi Murata, Predicting popularity trend in social media networks with multi-layer temporal graph neural network, 2024, doi: https://doi.org/10.1007/s40747-024-01402-6. DOI: https://doi.org/10.1007/s40747-024-01402-6

[21] Asosiasi Penyelenggara Jasa Internet Indonesia (APJII), Profil Internet Indonesia 2026. 2026.

[22] Wu, Y., Fu, C., He, Z. et al., Multi-view temporal graph neural network for numerous miniature cascade popularity prediction, 2025, doi: https://doi.org/10.1007/s11227-025-07484-4. DOI: https://doi.org/10.1007/s11227-025-07484-4

[23] Ning Xu, Xiaowen Wang, Jing Liu, Lanjun Wang, Xuanya Li, Mengxiao Zhu, Yongdong Zhang, An-An Liu, Model Can Be Subtle: Two Important Mechanisms for Social Media Popularity Prediction, 2025, doi: https://doi.org/10.1145/3705319. DOI: https://doi.org/10.1145/3705319

[24] Jana Gross, Full article: Influencer Marketing on Instagram: Empirical Research on Social Media Engagement with Sponsored Posts, 2022, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://www.tandfonline.com/doi/full/10.1080/15252019.2022.2123724

[25] I. C. Dang, C. M. C. Kwan, J. S. Jia, dan Y. Shi, When Words Meet Visuals: How Content Composition Drives Social Media Engagement for Marketer-Generated Content, J. Mark. Res., vol. 63, no. 1, hlm. 167190, Feb 2026, doi: 10.1177/00222437251373042. DOI: https://doi.org/10.1177/00222437251373042

[26] Simon zollo, Examining the relationship between content length and engagement with news outlets on multiple social media platforms - ScienceDirect, 2026, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://www.sciencedirect.com/science/article/pii/S0040162526001459 DOI: https://doi.org/10.1016/j.techfore.2026.124668

[27] Myoung-Jin Chae, Real-time marketing messages and consumer engagement in social media - ScienceDirect, 2025, Diakses: 11 Juni 2026. [Daring]. Tersedia pada: https://www.sciencedirect.com/science/article/abs/pii/S014829632500089X

[28] Z. Ye, J. Yang, F. Meng, M. Li, dan Y. Zhan, Integrating Temporal Interest Dynamics and Virality Factors for High-Precision Ranking in Big Data Recommendation, Electronics, vol. 14, no. 18, hlm. 3687, Jan 2025, doi: 10.3390/electronics14183687. DOI: https://doi.org/10.3390/electronics14183687

[29] J. Zhang dkk., THGFormer: Time-Aware Hypergraph Learning for Multimodal Social Media Popularity Prediction (Student Abstract), Proc. AAAI Conf. Artif. Intell., vol. 38, no. 21, hlm. 2370523706, Mar 2024, doi: 10.1609/aaai.v38i21.30534. DOI: https://doi.org/10.1609/aaai.v38i21.30534

[30] D. H. Kim, O. Kuru, J. Zeng, dan S. Kim, Fostering mask-wearing with virality metrics and social media literacy: evidence from the U.S. and Korea, Front. Psychol., vol. 14, Mei 2023, doi: 10.3389/fpsyg.2023.1151061. DOI: https://doi.org/10.3389/fpsyg.2023.1151061

[31] N. E. Beshet dkk., Trends Sentiment Unveiled Through Deep Dive into Social Media Data, Procedia Comput. Sci., vol. 275, hlm. 799808, Jan 2026, doi: 10.1016/j.procs.2026.01.092. DOI: https://doi.org/10.1016/j.procs.2026.01.092

[32] H. Peters, Y. Liu, F. Barbieri, R. A. Baten, S. C. Matz, dan M. W. Bos, Context-aware prediction of active and passiveuser engagement: Evidence from a large online social platform, J. Big Data, vol. 11, no. 1, hlm. 110, Agu 2024, doi: 10.1186/s40537-024-00955-0. DOI: https://doi.org/10.1186/s40537-024-00955-0

[33] L. Rivadeneira dan I. Loor, Evidential Reasoning Approach for Predicting Popularity of Instagram Posts, IEEE Access, vol. 12, hlm. 182603182617, 2024, doi: 10.1109/ACCESS.2024.3510637. DOI: https://doi.org/10.1109/ACCESS.2024.3510637