Analyzing Hybrid Deep Learning Models with Decomposition Methods: A Case Study of RSS3 Price in Thailand

Rubber, an economically important crop in Thailand, is cultivated and exported in various processed forms including ribbed smoked sheet (RSS), block rubber, mixed rubber, and liquid latex. To facilitate efficient trade planning and management of fluctuating rubber prices, accurate forecasting plays a crucial role as a supporting mechanism. At present, hybrid models based on decomposition methods have gained widespread adoption for this purpose. These decomposition and ensemble strategies can be categorized into two primary approaches: simultaneous forecasting and separate forecasting. In this research, our goal is to compare both approaches by utilizing Empirical Mode Decomposition (EMD) and Ensemble Empirical Mode Decomposition (EEMD) in conjunction with LSTM and GRU models. The analysis was carried out using data on the prices of the Thai RSS Grade 3.