Dynamic Decision Prediction Model based on GMARIMA and ETA Models

This paper presents the development and implementation of a dynamic decision prediction model that combines the Grey Prediction Model (GM), ARIMA, and Event Tree Analysis (ETA) to enhance the accuracy and reliability of predictions in complex systems. The methodology begins with data preprocessing using the GM model to handle small samples and limited information, followed by time series analysis using ARIMA to model and forecast based on historical trends. Event prediction is conducted through ETA, which evaluates potential future events and their impacts. The integrated model leverages the strengths of each component: GM is potent for cases with uncertain and limited data, ARIMA effectively captures temporal patterns, and ETA is proficient in event sequence analysis. The results show a significant improvement in forecasting accuracy and decision-making efficiency, particularly in areas such as energy consumption, financial market analysis, and infrastructure maintenance. Sensitivity and feasibility analyses confirm the model’s robustness and practicality, making it a valuable tool for system performance optimization, resource allocation, and strategic planning. This integrated approach provides a comprehensive framework for addressing complex prediction and decision-making challenges in dynamic environments.