Development of Social Work Demand Forecasting System Integrating Deep Learning and IoT Technology

The escalating demand for social work services, driven by global demographic shifts and socioeconomic challenges, underscores the necessity for innovative technological interventions to facilitate proactive service delivery. This research presents a comprehensive forecasting system for social work demand that integrates deep learning methodologies with Internet of Things (IoT) technology, aimed at enhancing both the efficiency and quality of services provided. The proposed system utilizes a hybrid model combining Convolutional Neural Networks (CNN) for spatial feature extraction and Long Short-Term Memory (LSTM) networks for temporal data modelling, achieving a prediction accuracy of 94.8%, which represents a 30.8% enhancement over conventional forecasting methods. The IoT framework consists of 5,320 sensing points that gather multimodal data from various sources, including environmental monitoring, health tracking, and safety alert systems. Implementation of this system across 20 social work institutions yielded notable improvements: a reduction in response time from 8,500 milliseconds to 320 milliseconds, an increase in system availability to 99.2%, and a user satisfaction rate of 91.5%. The cost-benefit analysis indicates a payback period of 1.7 years, with anticipated net benefits of NT$35.26 million over five years. This study offers a replicable technical framework for the modernization of social work practices while addressing critical ethical considerations such as privacy protection, algorithmic bias, and the dynamics of human-machine collaboration. The findings contribute to the advancement of more equitable and effective social welfare systems through the application of technological innovations.