Multi-Step Rubber Price Prediction Using Deep Learning Models with External Factors

The Thai rubber market plays a vital role in the nation's economy, holding a prominent position as one of the world's leading producers and exporters of natural rubber. It frequently faces challenges, including price volatility, driven by complex global supply and demand dynamics and numerous influencing factors. This study aimed to enhance multi-step price prediction models for ribbed smoked rubber sheets (RSS) through the application of advanced deep learning techniques. Various recurrent neural network architectures were examined, and the impact of incorporating external factors into the analysis was thoroughly assessed. The results demonstrated that multistep gated recurrent units coupled with top principal components of external features based on principal component analysis delivered superior performance, achieving an average RMSE of 3.54 and an average MAPE of 4.46% across 7-step predictions. A precise price estimation for key rubber products, including RSS, could enable informed decision-making for all stakeholders throughout the entire supply chain.