Financial Ratio Prediction for Accounting Information System Using Holt-Winters Exponential Smoothing

Abstract

Financial ratio analysis serves as a vital tool for assessing organizational performance and supporting decision-making processes. This study developed an accounting information system featuring financial ratio prediction capabilities powered by the Holt-Winters Exponential Smoothing algorithm—a method capable of handling time-series data characterized by trends and seasonality. Functional testing demonstrated that the system operates according to design, achieving a 94% success rate. Prediction results varied in accuracy: the current ratio achieved a Mean Absolute Percentage Error (MAPE) of 8.3%, indicating a favorable prediction outcome. Conversely, the Return on Investment (ROI), Return on Equity (ROE), and debt-to-equity ratios yielded MAPE values exceeding 20%, signaling a need for improved accuracy. These findings suggest that the system can serve as a relevant analytical tool for financial management; however, further development is required, particularly regarding algorithm parameter optimization and the provision of higher-quality data.