Crop Yield Prediction Using Machine Learning and Deep Learning Models

Abstract

In this project we are using machine learning and deep learning algorithms to predict future crop yield based on weather data such as temperature and rainfall. If farmers know the crop yield before sowing based on historical weather data, then he may take better decision. So, by employing machine/deep learning algorithms we can inform farmers about future crop yield. In proposed method we are using Irish Maize and Potato yield dataset to train all machine learning models and then these models can be used to predict future crop yield. In proposed method we are using random forest, SVR, DNN, CNN, ANN and LSTM. So, we have implemented all 6 algorithms on both datasets. To evaluate performance of each algorithm we are calculating MSE and R2 Score where MSE refers to mean square error (difference between TEST crop yield and predicted yield). R2 refers to correct prediction rate. So, for any algorithm MSE must be lower and R2 must be higher for better crop yield prediction.