Prediction of Weather Forecast for Smart Agriculture supported by Machine Learning
Francisco Raimundo, André Glória, Pedro Sebastião
Abstract
This paper introduces a study done to evaluate the use of machine learning regression techniques to predict the weather conditions of agricultural fields for smart irrigation systems. The proposed methodology is able to predict the temperature, precipitation, wind speed and evapotranspiration based on the field location and day. To discover the best model to achieve this, a set of machine learning techniques were implemented, including Linear Regression, Decision Tree, Random Forest and Neural Networks, being the results compared. Results shown that Random Forests and Decisions Trees achieve the best efficiency, after cross-validation. This paper includes a detailed description of the methodology, its implementation and the experimental results.