DATA DRIVEN HYBRID FRAMEWORK USING BILSTM AND GAN FOR PREDICTING MORINGA OLEIFERA GROWTH AND LEAF HEALTH
Keywords:
Moringa Oleifera, Generative Adversarial Networks, Bidirectional Long Short- Term Memory, Growth Prediction, Yield Estimation, Leaf Disease Detection, Precision Agriculture, Sustainable Farming.Abstract
Background: Sustainable agricultural productivity relies on precise forecasting of crop yield, growth dynamics, and early disease diagnosis. Moringa oleifera, a plant valued for its nutritional and therapeutic benefits, often faces inconsistent production due to unpredictable environmental variations and unrecognized leaf infections. Traditional prediction and detection models struggle to manage complex dependencies between climatic parameters and plant responses.
Methods: This research introduces a robust hybrid framework integrating Bidirectional Long Short-Term Memory (BiLSTM) networks with Generative Adversarial Networks (GAN) to address these challenges. The BiLSTM component effectively learns temporal dependencies from sequential agricultural data such as moisture content, sunlight exposure, and soil fertility indices, while the GAN component enriches the dataset by generating high-quality synthetic samples to enhance training diversity and reduce overfitting.
Results: Additionally, a deep learning-based leaf assessment module processes visual data to classify disease types and evaluate plant health status. The integrated model ensures precise yield forecasting, continuous growth monitoring, and early disease recognition through adaptive feature learning.
Conclusion: The hybrid BiLSTM–GAN framework achieved superior accuracy and generalization compared to existing methods, enabling better crop management and decision support in precision agriculture. This approach provides a scalable, data-driven pathway to strengthen sustainable Moringa oleifera cultivation and optimize agricultural resource utilization.
