Predicting Crop Yield with AI—A Comparative Study of DL and ML Approaches

Crop yield prediction plays a vital role in the agricultural industry as it directly impacts global food security and also crop yield prediction is essential in making informed decisions about crop selection what to grow and when to grow, production, marketing, managing risk, increasing productivity, ensuring food security and promoting environmental sustainability. Several machine learning and deep learning algorithms are widely applied to predict the yield of the crop. In this paper, we conducted a systematic literature review of various machine learning and deep learning-based studies to examine the methods and features used in crop yield prediction. Climate variables such as temperature (minimum and maximum), humidity, wind velocity, rainfall, sunlight hours, water level, soil content and soil type will have a great impact on crop yield. This study highlights that convolutional neural networks (CNN), followed by long-short term memory (LSTM) and deep neural networks (DNN) are the most commonly used DL algorithms for the yield of the crop. Deep learning algorithms achieved low prediction errors compared to other models.