INTEGRATING MULTILINGUAL TICKET CLASSIFICATION AND WORKLOAD FORECASTING FOR CAPACITY-AWARE ITSM ROUTING
DOI:
https://doi.org/10.18623/rvd.v23.5262Palavras-chave:
IT Service Management, Multilingual NLP, Workload Forecasting, Capacity-Aware Routing, Class ImbalanceResumo
Multinational organizations handle large volumes of multilingual IT Service Management (ITSM) tickets under time and capacity constraints. Centralized Service Desk routing can create delays and uneven workload distribution. While prior research on automated ticket classification emphasizes predictive accuracy, operational capacity limits are rarely incorporated into routing decisions. This study proposes and empirically evaluates a forecast-informed, capacity-aware routing framework integrating multilingual DistilBERT classification with XGBoost-based workload forecasting. Using historical IT tickets from an Indonesian multinational firm, routing performance on a held-out evaluation set reached 96% accuracy and a macro F1 score of 0.87. A controlled stress test demonstrated measurable trade-offs between routing accuracy and workload stability, showing how routing decisions perform under real operational constraints.
Referências
Ahmad, W. U., Li, H., Chang, K.-W., & Mehdad, Y. (2021). Syntax-augmented multilingual BERT for cross-lingual transfer. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics (ACL 2021). Association for Computational Linguistics. https://aclanthology.org/2021.acl-long.350.pdf
Cui, X. (2025). Addressing data imbalance in transformer-based multi-label emotion detection with weighted loss (arXiv:2507.11384). arXiv. https://arxiv.org/abs/2507.11384
Deshpande, A., Talukdar, P., & Narasimhan, K. (2022). When is BERT multilingual? Isolating crucial ingredients for cross-lingual transfer. arXiv. https://arxiv.org/abs/2110.14782
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019). BERT: Pre-training of deep bidirectional transformers for language understanding. In Proceedings of NAACL-HLT 2019 (pp. 4171–4186). Association for Computational Linguistics. https://doi.org/10.18653/v1/N19-1423
Eichenseer, P., Hans, L., & Winkler, H. (2025). A data-driven machine learning model for forecasting delivery positions in logistics for workforce planning. Supply Chain Analytics, 9, 100099. https://doi.org/10.1016/j.sca.2024.100099
Gardazi, N. M., Daud, A., Malik, M. K., et al. (2025). BERT applications in natural language processing: A review. Artificial Intelligence Review, 58, 166. https://doi.org/10.1007/s10462-025-11162-5
Henning, S., Beluch, W., Fraser, A., & Friedrich, A. (2023). A survey of methods for addressing class imbalance in deep-learning based NLP. In Proceedings of EACL 2023 (pp. 523–540). https://aclanthology.org/2023.eacl-main.38
Jung, V., & van der Plas, L. (2024). Understanding the effects of language-specific class imbalance in multilingual fine-tuning. In Findings of ACL: EACL 2024 (pp. 2368–2376). Association for Computational Linguistics. https://aclanthology.org/2024.findings-eacl.157
Liu, Z., Benge, C., & Jiang, S. (2023). Ticket-BERT: Labeling incident management tickets with language models. arXiv. https://doi.org/10.48550/arXiv.2307.00108
Mandal, N., Malhotra, S., Agarwal, A., Ray, A., & Sridhara, G. (2019). Automated dispatch of helpdesk email tickets: Pushing the limits with AI. Proceedings of the AAAI Conference on Artificial Intelligence, 33(1), 9381–9388. https://doi.org/10.1609/aaai.v33i01.33019381
Marcuzzo, M., Zangari, A., Giudice, L., Gasparetto, A., Schiavinato, M., & Albarelli, A. (2022). A multi-level approach for hierarchical ticket classification. In Proceedings of the 8th Workshop on Noisy User-generated Text (W-NUT 2022) (pp. 201–214). Association for Computational Linguistics.
Osma-Valenzuela, A. J., & Torrado-Castro, D. J. (2024). Predictive model for IT capacity management based on machine learning. Journal of Computer and Electronic Science: Theory and Applications, 5(1), 35–49. https://revistascientificas.cuc.edu.co/CESTA/article/view/5956
Petridis, C. (2024). Text classification: Neural networks vs machine learning models vs pre-trained models. arXiv. https://doi.org/10.48550/arXiv.2412.21022
Sanh, V., Debut, L., Chaumond, J., & Wolf, T. (2019). DistilBERT: A distilled version of BERT. arXiv. https://doi.org/10.48550/arXiv.1910.01108
Serrano, J., Faustino, J., Adriano, D., Pereira, R., & da Silva, M. M. (2021). An IT service management literature review: Challenges, benefits, opportunities and implementation practices. Information, 12(3), 111. https://doi.org/10.3390/info12030111
Subbarao, M. V., Venkatarao, K., & Suresh, C. H. (2022). Automation of incident response and IT ticket management by ML and NLP mechanisms. Journal of Theoretical and Applied Information Technology, 100(12), 3945–3951. https://www.jatit.org/volumes/Vol100No12/13Vol100No12.pdf
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, Ł., & Polosukhin, I. (2017). Attention is all you need. Advances in Neural Information Processing Systems, 30, 5998–6008.
Yıldırım, Ş., Yücekaya, A. D., Hekimoğlu, M., Ucal, M., Aydın, M. N., & Kalafat, İ. (2023). AI-driven predictive maintenance for workforce and service optimization in the automotive sector. Applied Sciences, 13(20), 11111. https://doi.org/10.3390/app15116282
Zhang, L., Bian, W., Qu, W., Tuo, L., & Wang, Y. (2021). Time series forecast of sales volume based on XGBoost. Journal of Physics: Conference Series, 1873(1), 012067. https://doi.org/10.1088/1742-6596/1873/1/012067
Downloads
Publicado
Como Citar
Edição
Seção
Licença
Submeto (emos) o presente trabalho, texto original e inédito, de minha (nossa) autoria, à avaliação de Veredas do Direito - Revista de Direito, e concordo (amos) que os direitos autorais a ele referentes se tornem propriedade exclusiva da Revista Veredas, sendo vedada qualquer reprodução total ou parcial, em qualquer outra parte ou outro meio de divulgação impresso ou eletrônico, dissociado de Veredas do Direito, sem que a necessária e prévia autorização seja solicitada por escrito e obtida junto ao Editor-gerente. Declaro (amos) ainda que não existe conflito de interesse entre o tema abordado, o (s) autor (es) e empresas, instituições ou indivíduos.
Reconheço (Reconhecemos) ainda que Veredas está licenciada sob uma LICENÇA CREATIVE COMMONS:
Licença Creative Commons Attribution 3.0
