<![CDATA[Financial literacy and personal cash flow management are critical components of individual economic stability. However, the volatility of daily transaction patterns often complicates manual monitoring, leading to liquidity deficit risks. This study aims to develop a Decision Support Model for personal cash flow monitoring by comparing the effectiveness of AutoRegressive Integrated Moving Average (ARIMA) and Holt-Winters Exponential Smoothing methods. Using the ANZ Virtual Internship transaction dataset, this research processes historical customer daily balance data through preprocessing, stationarity testing, and predictive modeling stages. Performance evaluation results demonstrate that the Holt-Winters method is more accurate in capturing customer balance trends with a Mean Absolute Error (MAE) of AUD 476.80 and a Root Mean Square Error (RMSE) of AUD 568.96, outperforming ARIMA which yielded higher error rates. Based on these results, the Holt-Winters algorithm is integrated into an Early Warning System (EWS) logic that maps balance predictions into three risk zones: Safe, Warning, and Critical. The proposed model is capable of providing future financial health status visualization as a risk mitigation instrument for banking customers]]>
Copyrights © 2026
