<![CDATA[Cash flow forecasting is important because changes in cash availability can affect liquidity management, funding decisions, and operational stability. This study aims to develop and evaluate an interpretable predictive model for next-period Net Cash Flow using machine learning Linear Regression. Quarterly panel data were collected from 11 multifinance issuers listed on the Indonesia Stock Exchange during 2021–2025. The predictors include Operating Cash Flow, Investing Cash Flow, Financing Cash Flow, and company identity, while the target is next-period Net Cash Flow. Microsoft Excel was used for initial data preparation, checking, and verification, while Google Colab with Python, Pandas, NumPy, scikit-learn, and Matplotlib was used for data processing, modeling, evaluation, charts, and dashboard visualization. From 220 raw observations, 209 model-ready observations were obtained, consisting of 176 training and 33 testing observations. Linear Regression produced MAE of 5,175,010,760.53 and RMSE of 11,439,892,766.99, lower than naive forecasting with MAE of 7,569,470,101.77 and RMSE of 16,555,565,343.09. However, the negative testing R-squared indicates limited performance under heterogeneous and extreme cash-flow movements. These findings show that Linear Regression can provide a measurable and interpretable initial basis for cash flow prediction, although broader data, additional variables, and stronger models are still required.]]>
Copyrights © 2026
