Forecasting Educational Inequality in China for Sustainable Development: A Hybrid Framework of GM(1,1) and CS-SVR

Educational equality is essential for achieving social justice and sustainable development. Accurately predicting the trend of educational inequality is important for improving education systems and ensuring equitable resource allocation. In this paper, the Educational Gini (E-Gini) index is calculated based on the population aged 6 and above in China from 2002 to 2024, quantifying educational inequality. To forecast the future trend in the E-Gini index, a hybrid prediction framework based on the grey prediction model (GM(1,1)) and Cuckoo search-support vector regression (CS-SVR) model is proposed. This framework incorporates three influencing factors, including government budget spending on education, per capita consumption expenditure on education, and the Consumer Price Index (CPI) for education. The results show that the E-Gini of China generally declines from 2002 to 2024 with fluctuations. The proposed approach predicts the E-Gini value of 2024 as 0.220130, while the actual value is 0.2206, corresponding to an absolute error of 0.000470 and a relative error of 0.213%. In the benchmark comparison, the proposed model outperforms the linear trend model, the univariate GM(1,1), the naive persistence model, ARIMA, and the standard SVR model. The comparative analysis demonstrates that the proposed framework effectively captures the inherent patterns of educational inequality and reveals its trends. The proposed framework serves as a valuable tool for forecasting trends in educational inequality and informing policy decisions.