<![CDATA[This study aims to determine the best forecasting method for CAR by comparing three approaches, namely automatic autoregressive integrated moving average (auto-ARIMA), multilayer perceptron (MLP) neural networks, and ensemble method and to see their relationship to resilience, prudence, public trust, and stability of Islamic banking institutions. The CAR data used is monthly time series data published by the Financial Services Authority (OJK) in Indonesia for the period 2015-2025. Training and testing data are used to evaluate forecasting performance using mean absolute error (MAE), mean squared error (MSE), and mean absolute percentage error (MAPE). Forecasting results using the auto-ARIMA (0,1,0) model, the best method, confirmed that CAR is on a stable and sustainable path. This finding reinforces CAR's role as a multidimensional indicator linking financial performance, institutional resilience, sharia compliance, prudence, and social responsibility of Islamic banks to the community. The accuracy of this forecasting has direct implications for strengthening Islamic banking governance by increasing capital resilience in the face of future economic shocks. The ability to accurately predict CAR allows management to prioritize prudent principles in financing distribution, thereby mitigating the risk of systemic failure. Furthermore, well-planned capital ratio stability will strengthen public confidence in the security of funds in Islamic financial institutions. Reliable CAR forecasting not only supports managerial decision-making, but also contributes to strengthening the stability and credibility of Islamic banking as an institution responsible for society and the economy in Indonesia.]]>
Copyrights © 2026
