A Machine Learning Framework for Optimizing Public Service Delivery on Social Media in Oman

This research proposes an AI based smart guide framework to optimize public service dissemination via social media for Oman's government. Current systems lack structured AI tools for citizen interaction, leading to inefficiencies in service delivery. Hence a two-phase framework using Natural Language Processing (NLP) and machine learning is designed. Phase 1 implements a case study at UTAS, where students and parents submit service requests via social media; Phase 2 extends the system to government agencies. The framework includes pre-processing like data aggregation, noise removal, followed by text mining for keyword extraction and intent classification. Further, AI-based processing for similarity matching and automated responses are conducted. Preliminary data analysis showed approximately 85% accuracy in request categorization. The system reduces response time and improves citizen satisfaction by automating 70% of routine queries. This work demonstrates how AI-enhanced social media analytics can bridge gaps in government-citizen communication. Future work includes IoT integration for smart city scalability.