TRIPLE EXPONENTIAL SMOOTHING AND RANDOM FOREST ALGORITHMS FOR A HYBRID RISK FORECASTING BY USING WEIGHTED EVENT DATA
Risk forecasting plays a crucial role in various fields, ranging from finance to environmental management. This research proposes a novel approach that combines the Triple Exponential Smoothing (TES) method and the Random Forest (RF) algorithm to ach...
TRIPLE EXPONENTIAL SMOOTHING AND RANDOM FOREST ALGORITHMS FOR A HYBRID RISK FORECASTING BY USING WEIGHTED EVENT DATA
Risk forecasting plays a crucial role in various fields, ranging from finance to environmental management. This research proposes a novel approach that combines the Triple Exponential Smoothing (TES) method and the Random Forest (RF) algorithm to achieve more accurate and reliable risk forecasting. TES is highly effective in predicting seasonality and trends in time-series data, while RF excels in handling complex non-linear relationships and mitigating model overfitting. By integrating these two methods, the proposed hybrid model aims to leverage the strengths of each method to enhance the accuracy and performance of risk prediction. This research represents an effort to improve the performance of the seventh step in the Improved Multi-Attribute Risk Assessment (IMARA) method, which is risk forecasting. The experiments in this study utilize a dataset from PT Terminal Peti Kemas Surabaya to evaluate the performance of the proposed approach. The results of this research are expected to demonstrate an improvement in the accuracy and performance of risk prediction compared to the risk prediction performance generated by the TES and RF methods when applied individually, especially in dealing with complex data patterns and dynamic risk factors. This research uses three scenarios for combining TES and RF to obtain the highest accuracy in risk forecasting. The results from the first phase of the experiments showed that Scenario 3 yields slightly better accuracy compared to the two other scenarios.
