International Journal of Modern Education and Computer Science (IJMECS)
IJMECS Vol. 18, No. 5, 8 Oct. 2026
Cover page and Table of Contents: PDF (size: 1899KB)
Citizens as Sensors: Fusing Multilingual Feedback with System Logs for Predictive Deployment Monitoring in Resource-Constrained Public Services
PDF (1899KB), PP.192-212
Views: 0 Downloads: 0
Author(s)
Johnson Nuviadenu
1,*
Themba Masombuka
1
Ernest Mnkandla
1
Malusi Sibiya
1
Index Terms
Deployment Monitoring, Multilingual NLP, Public Digital Services, Predictive Analytics, Citizen Feedback, Resource-Constrained Environments.
Abstract
Public digital services in developing economies have become central to citizens’ access to government functions, yet frequent deployment failures undermine trust and service delivery. This study investigated whether multilingual citizen feedback could serve as an early-warning signal for deployment failures in resource-constrained public sector digital services. A bimodal failure detection framework was developed combining English system logs with citizen feedback in Akan, Ewe and Ga languages from a national citizen service portal in Ghana handling over 50,000 monthly users. Rather than building language-specific classifiers requiring annotated training data, the study leveraged instruction-tuned large language models as general-purpose interpreters using few-shot prompting. Analysis of 18 months of deployment logs and 1,217 citizen feedback messages revealed that a fusion model combining XGBoost for log analysis and prompt-engineered Qwen 2.5-3B-Instruct for feedback interpretation achieved 82 per cent recall and 76 per cent precision, representing a 21 percentage point improvement over log-only approaches. Citizen feedback preceded technical alerts in 32 per cent of failure cases, with an average detection advantage of 2.3 hours, which increased to 3.8 hours during peak usage periods. This study provides empirical evidence that instruction-tuned large language models can extract operationally useful failure signals from low-resource languages and improve post-deployment early warning when fused with conventional log-based monitoring. The findings imply that public institutions in multilingual developing contexts can adopt LLM-based monitoring without requiring annotated datasets or language-specific processing pipelines.
Cite This Paper
Johnson Nuviadenu, Themba Masombuka, Ernest Mnkandla, Malusi Sibiya, "Citizens as Sensors: Fusing Multilingual Feedback with System Logs for Predictive Deployment Monitoring in Resource-Constrained Public Services", International Journal of Modern Education and Computer Science(IJMECS), Vol.18, No.5, pp. 192-212, 2026. DOI:10.5815/ijmecs.2026.05.11
Reference
[1]United Nations Department of Economic and Social Affairs, “E-Government Survey 2020: Digital Government in the Decade of Action for Sustainable Development,” United Nations, New York, NY, USA, Rep. ST/ESA/PAD/SER.E/214, 2020.
[2]United Nations Conference on Trade and Development, “Digital Economy Report 2021: Cross-border Data Flows and Development: For Whom the Data Flow,” UNCTAD, Geneva, Switzerland, Rep. UNCTAD/DER/2021, 2021.
[3]R. Heeks, “Information systems and developing countries: Failure, success, and local improvisations,” The Information Society, vol. 18, no. 2, pp. 101–112, 2002, doi: 10.1080/01972240290075039.
[4]J. Nuviadenu, T. Masombuka, E. Mnkandla, and M. Sibiya, “Machine learning for software deployment in the public sector: A systematic review and research agenda for African contexts,” Journal of Information Systems and Informatics, vol. 8, no. 4, pp. 5241-5290, Aug. 2026. https://doi.org/10.63158/journalisi.v8i4.1839
[5]M. Chen, A. X. Zheng, J. Lloyd, M. I. Jordan, and E. Brewer, “Failure diagnosis using decision trees,” in Proceedings of the International Conference on Autonomic Computing, New York, NY, USA, 2004, pp. 36–43, doi: 10.1109/ICAC.2004.1301345.
[6]V. Le and H. Zhang, “Log-based anomaly detection with deep learning,” in Proceedings of the 44th International Conference on Software Engineering, 2022, pp. 1356–1367, doi: 10.1145/3510003.3510155.
[7]B. Naznin, Z. Quayyum, J. Tajree, D. Barua, M. Ahsan, F. Kabir, et al., “Assessment of public-private partnership (PPP) models in health systems in least developed, low income and lower-middle-income countries and territories: A protocol for a systematic review,” 2022, doi: 10.21203/rs.3.rs-1399451/v1.
[8]Trist E. and Bamforth K. W. “Some social and psychological consequences of the longwall method of coal-getting,” Human Relations, 1951;4(1):3-38. doi: 10.1177/001872675100400101
[9]L. E. Lwakatare, I. Crnkovic´, and J. Bosch, “DevOps for AI – Challenges in development of AI-enabled applications,” in Proceedings of the 2020 International Conference on Software, Telecommunications and Computer Networks (SoftCOM), 2020, pp. 1–6, doi: 10.23919/softcom50211.2020.9238323.
[10]J. R. Valdivia, J. Gamboa-Cruzado, and P. D. l. C. V. d. Villa, “Systematic literature review on machine learning and its impact on APIs deployment,” Computacio´n y Sistemas, vol. 27, no. 4, 2023, doi: 10.13053/cys-27-4-4371.
[11]P. Liang, Y. Wu, Z. Xu, S. Xiao, and J. Yuan, “Enhancing security in DevOps by integrating artificial intelligence and machine learning,” Journal of Theory and Practice of Engineering Science, vol. 4, no. 2, pp. 31–37, 2024, doi: 10.53469/jtpes.2024.04(02).05.
[12]M. D. Syer, Z. M. Jiang, M. Nagappan, A. E. Hassan, M. Nasser, and P. Flora, “Leveraging performance counters and execution logs to diagnose memory-related performance issues,” in Proceedings of the 2013 IEEE International Conference on Software Maintenance, Eindhoven, Netherlands, 2013, pp. 110–119, doi: 10.1109/ICSM.2013.22.
[13]F. Rahman and P. Devanbu, “How, and why, process metrics are better,” in Proceedings of the 2013 35th International Conference on Software Engineering (ICSE), San Francisco, CA, USA, 2013, pp. 432–441, doi: 10.1109/ICSE.2013.6606589.
[14]Q. N. Nkohkwo and M. S. Islam, “Challenges to the successful implementation of e-government initiatives in sub-Saharan Africa: A literature review,” Electronic Journal of e-Government, vol. 11, no. 2, pp. 253– 267, 2013. [Online]. Available: https://academic-publishing.org/index.php/ejeg/article/view/585. [Accessed: Oct. 11, 2025]
[15]J. D. Twizeyimana and A. Andersson, “The public value of e-government – A literature review,” Government Information Quarterly, vol. 36, no. 2, pp. 167–178, Apr. 2019, doi: 10.1016/j.giq.2019.01.001.
[16]S. Kim, H. J. Kim, and H. Lee, “An institutional analysis of an e-government system for anti-corruption: The case of OPEN,” Government Information Quarterly, vol. 26, no. 1, pp. 42–50, 2009, doi: 10.1016/j.giq.2008.09.002.
[17]L. F. Luna-Reyes and J. R. Gil-Garc´ıa, “Digital government transformation and internet portals: The co-evolution of technology, organizations, and institutions,” Government Information Quarterly, vol. 31, no. 4, pp. 545–555, 2014, doi: 10.1016/j.giq.2014.08.001.
[18]N. Ude, K. Ude, U. Ugbor, C. Igwe, and E. Ogu, “E-governance and economic development in sub-Saharan Africa: A case of Nigeria,” International Journal of Development Strategies in Humanities, Management and Social Sciences, vol. 11, no. 1, pp. 87–100, 2021, doi: 10.48028/iiprds/ijdshmss.v11.i1.07.
[19]T. Schuppan, “E-government in developing countries: Experiences from sub-Saharan Africa,” Government Information Quarterly, vol. 26, no. 1, pp. 118–127, 2009, doi: 10.1016/j.giq.2008.01.006.
[20]Masakhane, “Masakhane: A grassroots NLP community for Africa, by Africans,” Masakhane, 2024. [Online]. Available: https://www.masakhane.io/. [Accessed: Oct. 11, 2025].
[21]A. L. Tonja, B. F. Dossou, J. Ojo, J. Rajab, F. Thior, E. P. Wairagala, A. Anuoluwapo, P. Moiloa, J. Abbott, V. Marivate, and B. Rosman, “InkubaLM: A small language model for low-resource African languages,” arXiv preprint arXiv:2408.17024, 2024. [Online]. Available: https://arxiv.org/abs/2408.17024. [Accessed: Oct. 19, 2025]
[22]W. Nekoto, V. Marivate, T. Matsila, T. Fasubaa, T. Fagbohungbe, S. Akinola, et al., “Participatory research for lowresourced machine translation: A case study in African languages,” in Findings of the Association for Computational Linguistics: EMNLP 2020, 2020, doi: 10.18653/v1/2020.findings-emnlp.195.
[23]K. Ogueji, Y. Zhu, and J. Lin, “Small data? No problem! Exploring the viability of pretrained multilingual language models for low-resourced languages,” in Proceedings of the 1st Workshop on Multilingual Representation Learning, 2021, doi: 10.18653/v1/2021.mrl-1.11.
[24]T. B. Brown, B. Mann, N. Ryder, M. Subbiah, J. Kaplan, P. Dhariwal, et al., “Language models are few-shot learners,” in Proceedings of the 34th International Conference on Neural Information Processing Systems, Vancouver, BC, Canada, 2020. [Online]. Available: https://dl.acm.org/doi/abs/10.5555/3495724.3495883. [Accessed: Nov. 1, 2025]
[25]J. Wei, M. Bosma, V. Y. Zhao, K. Guu, A. W. Yu, B. Lester, N. Du, A. M. Dai, and Q. V. Le, “Finetuned language models are zero-shot learners,” in Proc. 10th Int. Conf. on Learning Representations (ICLR), 2022.
[26]R. K. Logan, I. Balazevic, E. S. Wallace, F. Petroni, S. Singh, and S. Riedel, “Cutting down on prompts and parameters: Simple few-shot learning with language models,” in Findings of the Association for Computational Linguistics: ACL 2022, 2022, doi: 10.18653/v1/2022.findings-acl.222.
[27]S. Raschka, “Understanding and coding self-attention, multi-head attention, causal-attention, and cross-attention in LLMs,” Sebastian Raschka AI Research Magazine, Jan. 14, 2024. [Online]. Available: https://magazine. sebastianraschka.com/p/understanding-and-coding-self-attention. [Accessed: Nov. 11, 2025].
[28]E. A. Khadem and A. Movaghar, “From challenges to metrics: An LLM-driven DevOps recommendation system grounded in evidence-based mappings,” Array, vol. 28, p. 100547, 2025, doi: 10.1016/j.array.2025.100547.
[29]A. Fan et al., “Large language models for software engineering: Survey and open problems,” in Proceedings of the 2023 IEEE/ACM International Conference on Software Engineering: Future of Software Engineering (ICSE-FoSE), Melbourne, Australia, 2023, pp. 31–53, doi: 10.1109/ICSE-FoSE59343.2023.00008.
[30]A. Meijer and M. Thaens, “Alignment 2.0: Strategic use of new internet technologies in government,” Government Information Quarterly, vol. 27, no. 2, pp. 113–121, 2010, doi: 10.1016/j.giq.2009.12.001.
[31]R. Medaglia, “eParticipation research: Moving characterization forward (2006–2011),” Government Information Quarterly, vol. 29, no. 3, pp. 346–360, 2012, doi: 10.1016/j.giq.2012.02.010.
[32]G. A. Porumbescu, “Linking public sector social media and e-government website use to trust in government,” Government Information Quarterly, vol. 33, no. 2, pp. 291–304, 2016, doi: 10.1016/j.giq.2016.04.006.
[33]A. Yang et al., “Qwen2.5 Technical Report,” arXiv preprint arXiv:2412.15115, May 2025. [Online]. Available: https://doi.org/10.48550/arXiv.2412.15115. [Accessed: Nov. 1, 2025].
[34]A. Q. Jiang et al., “Mistral 7B,” arXiv preprint arXiv:2310.06825, Oct. 2023. [Online]. Available: https:// arxiv.org/abs/2310.06825. [Accessed: Nov. 3, 2025].
[35]T. Chen and C. Guestrin, “XGBoost: A scalable tree boosting system,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, pp. 785–794, doi: 10.1145/2939672.2939785.
[36]Y. Zhou, A. I. Muresanu, Z. Han, K. Paster, S. Pitis, H. Chan, and J. Ba, “Large language models are human-level prompt engineers,” arXiv preprint arXiv:2211.01910, Nov. 2022. [Online]. Available: https://arxiv.org/ abs/2211.01910. [Accessed: Nov. 11, 2025].
[37]P. Lewis, E. Perez, A. Piktus, F. Petroni, V. Karpukhin, N. Goyal, H. Ku¨ttler, M. Lewis, W. Yih, T. Rockta¨schel, S. Riedel, and D. Kiela, “Retrieval-augmented generation for knowledge-intensive NLP tasks,” in Advances in Neural Information Processing Systems (NeurIPS), vol. 33, 2020, pp. 9459–9474. [Online]. Available: https://dl. acm.org/doi/abs/10.5555/3495724.3496517. [Accessed: Nov. 8, 2025]
[38]Z. Jiang, J. Liu, Y. Li, H. Huang, X. He, T. Zhang, J. Chen, Y. Li, R. Shi, and M. R. Lyu, “LogPilot: Intent-aware and scalable alert diagnosis for large-scale online service systems,” in Proc. 2025 40th IEEE/ACM Int. Conf. on Automated Software Engineering (ASE), 2025, pp. 3299–3310, doi: 10.1109/ASE63991.2025.00272.
[39]W. Xu, J. Luo, T. Huang, K. Sui, J. Geng, Q. Ma, I. Akasaka, X. Shi, J. Tang, and P. Cai, “LogSage: An LLM-based framework for CI/CD failure detection and remediation with industrial validation,” in Proc. 2025 40th IEEE/ACM Int. Conf. on Automated Software Engineering (ASE), 2025, pp. 3742–3753, doi: 10.1109/ASE63991.2025.00310.
[40]Y. Liu, J. Cheng, J. Li, F. Tsung, H. Yin, and T. He, “A pre-training and adaptive fine-tuning framework for graph anomaly detection,” arXiv preprint arXiv:2504.14250, 2025. [Online]. Available: https://arxiv.org/abs/ 2504.14250
[41]B. Cohen, E. Khwaja, K. Wang, C. Masson, E. Rame´, Y. Doubli, and O. Abou-Amal, “Toto: Time series optimized transformer for observability,” arXiv preprint arXiv:2407.07874, 2024. [Online]. Available: https://arxiv. org/abs/2407.07874