Rainfall Prediction: Analysis of Machine Learning Algorithms and Ensemble Techniques

Abstract

In 2018, approximately 60% of land in India was reported as Agricultural Farmland. Most of the farmers in India depend on rainfall as the primary source of irrigation which implies that rainfall is directly linked to the yield. The quantity of rain a land receives helps in planning which crop to sow and also in which month the farmer should begin farming. Precipitation is not only limited to agriculture, even the infrastructure development sector has to keep track of the monsoon season and amount of rainfall that a construction site will receive because of the influence it has on construction projects. It also serves as one of the most important sources of freshwater for all living things on the planet. As rainfall prediction model gives data on the impact of various climatological variables on rainfall amounts it has many applications across many sectors. We propose a study with analysis of various machine learning algorithms for rainfall prediction. An accuracy of 95% was obtained with XGBoost, which is a gradient boosting framework. This also serves to inform the robusticity of the proposed model in comparison.