Generative Modeling Techniques for Simulating Rare Agricultural Events on Prediction of Wheat Yield Production

Wheat yield prediction is key to food security and optimal resource use. In this paper, Generative AI is use for specifically GANs and VAEs, to predict wheat yield under different weather conditions, uses the crop yield prediction dataset from Kaggle which has temperature, rainfall and soil quality as parameters and build models that beat traditional statistical methods. By leveraging advanced AI models, such as Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), the model generates synthetic data that mimics these rare occurrences. The GAN generates synthetic data that enriches the training set and improves prediction accuracy. The VAE captures the complex non-linear relationships between input features and output. The combination of both models gives us a robust framework for yield prediction. Experimental results shows that the GAN dataset gives 14% improvement over baseline models. VAE model gives 13% improvement on its own, showing its ability to handle the variability and complexity of agricultural data. This dual model approach is a big leap in agricultural analytics and will enable more robust and adaptive farming in the face of climate change.