Spatiotemporal evolution and forecasting of public attention to special education: a GIS and SARIMA-LSTM based analysis

Understanding public attention to special education is vital for promoting educational equity and achieving the Sustainable Development Goals (SDGs). This study aims to examine the spatiotemporal patterns of public attention to special education in China and to develop and evaluate a hybrid forecasting model to anticipate trends in such attention. Temporal trends and spatial disparities were assessed using geographic information system (GIS) techniques and spatial autocorrelation analysis. A hybrid forecasting approach combining Seasonal Autoregressive Integrated Moving Average (SARIMA) and long short-term memory (LSTM) models was applied, and its accuracy was evaluated using multiple performance metrics. The results suggest that public attention to special education remains stronger in the developed eastern regions and lags in the west. There are clear regional disparities and seasonal peaks in public attention, with the hybrid model outperforming single-model approaches in forecasting accuracy. This study integrates big data analytics, spatiotemporal modeling, and hybrid forecasting, contributing to future assessments of public attention to special education. The findings illustrate a framework for effectively improving responsiveness to educational policies, enhancing the efficiency of educational resource allocation, and promoting equitable and sustainable development in education.