AN INTELLIGENT APPROACH FOR CROP WATER FOOTPRINT PREDICTION

Abstract

Agriculture is the largest consumer of water; enhancement of the water level irrigation is essential for sustainability. This project employs the Random Forest Regressor for crop specifications as per hectare with considering the features such as crop type, seasonal data, location and meteorological data. To improve the robustness of the model performance, data preprocessing, Feature Engineering and Exploratory Data Analysis are used. The trained model is incorporated with a Flask Based web application, enabling the user, farmer, researchers and policymakers to custom their inputs and obtain their regional and crop specific predictions of water footprint. An in- built water calculator helps in manual estimations of predicting the water level required by specific crops along with yield area in cubic meters. By the combination of Machine Learning with user interface, it helps in the prediction of water footprint by considering the different features and improving the water conservation.