TabNet for crop yield prediction: promoting sustainability and climate resilience in Saudi Arabian agriculture

Predicting crop yield is essential for managing agricultural production and addressing sustainability in the face of climate change. Rising temperatures and extreme weather events coupled with agriculture’s contribution to greenhouse gas emissions underscore the need for data-driven solutions. This paper proposes a crop production prediction system using an Artificial Neural Network (ANN)-based TabNet model trained on extensive crop datasets. The model delivers actionable insights into optimal crop rotation strategies, particularly suited to Saudi Arabia’s agricultural conditions. TabNet demonstrated outperform prior models by achieving R2 scores of 0.9799 on Saudi Arabia’s Crop Yield dataset and 0.9850 on a benchmark dataset. The recommendations aim to enhance crop productivity, improve soil health, and mitigate climate change effects. By integrating ANN into agricultural decision-making, this approach supports sustainable farming practices and climate resilience.