Abstract

Accurate forecasting of Agile program performance is essential for effective project planning, risk management, and resource optimization in dynamic organizational environments. Traditional statistical and rule-based forecasting techniques often struggle to capture the nonlinear and seasonal patterns inherent in Agile program data. To address this limitation, this study proposes a Bidirectional Hybrid Long Short-Term Memory (LSTM) model for trend analysis and forecasting of Agile program performance. The model was developed using three years of synthetic Agile revenue data enhanced with trend and seasonal variations to simulate real-world performance fluctuations. The bidirectional structure enables the model to learn temporal dependencies in both forward and backward directions, while the hybrid dense layers enhance its capacity to model nonlinear relationships. The experimental results demonstrate that the proposed model achieved a Mean Absolute Error (MAE) of 2,483.43, a Root Mean Square Error (RMSE) of 2,923.07, and a Mean Absolute Percentage Error (MAPE) of 6.63 percent, indicating high predictive accuracy and stability. The forecasting outputs closely followed the actual revenue trends within a ±5 percent confidence interval, effectively capturing both mid-year performance declines and year-end growth patterns. These findings confirm that the Bidirectional Hybrid LSTM model provides an accurate and robust framework for forecasting Agile program performance. The model’s conservative prediction tendency and strong generalization ability make it a valuable decision-support tool for Agile management, enabling organizations to anticipate performance changes, allocate resources efficiently, and improve overall project predictability.

Keywords: Bidirectional LSTM, Agile Forecasting, Deep Learning, Time-Series, Decision Support

How to Cite:

Henderi, H. & Sofiana, S., (2026) “Deep Learning-based Trend Analysis and Forecasting of Agile Program Performance using LSTM Neural Networks: A Data-Driven Decision Support Approach”, Agile Management 1(4), 272-285. doi: https://doi.org/10.63913/ail.v1i4.141