Spatiotemporal Hybrid Random Forest Model for Tea Yield Prediction Using Satellite-Derived Variables

Crop yield forecasting is critical for enhancing food security and ensuring an appropriate supply. It to complete this activity with high precision at the regional national levels facilitate speedy decision-making. Tea a big cash crop that contributes significantly economic development, market of USD 200 billion in 2020 expected reach over 318 by 2025. As developing country, Bangladesh can be greater part industry increase its exports through tea production favorable climatic features land quality. Regrettably, has not increased since 2008 like many other countries, despite having suitable conditions, which why quantifying imperative. This study developed novel spatiotemporal hybrid DRS–RF model dragonfly optimization (DR) algorithm support vector regression (S) as feature selection approach. used satellite-derived hydro-meteorological variables between 1981 from twenty stations across address dependency predictor (Y). The results illustrated proposed improved standalone machine learning approaches, least relative error value (11%). indicates integrating random forest SVR-based improves prediction performance. approach help combat risk management countries.