Rainfall Prediction Using Catboost Machine Learning Algorithm

Forecasting rainfall is a complex task in weather prediction, as it plays a crucial role in managing reservoir water levels, directly affecting the water resource available to us. Climate change has made rainfall amounts unpredictable, which can lead to either overflow or drying of reservoirs. Monitoring changing rainfall patterns is essential to understand the impact on climate change, agriculture, and daily life. Accurate rainfall forecasting is crucial for countries like India, which heavily rely on cultivation. However, due to the vibrant type of the environment, statistical practices often decline to give accurate rainfall predictions. Global warming, a pressing concern that affects humankind, is accelerating climate change, causing oceans and the atmosphere to warm, and leading to floods and droughts. Uneven rainfall/precipitation is one of the foremost effects of such phenomenon. Precipitation estimating has become tedious and is being counted by main world agencies. Precipitation is a critical climatical feature that influence being activities like farming manufacture, construction, energy production, and tourism. Hence, accurate rainfall forecasting is essential to mitigate these effects. Various methods exist for predicting rainfall, but this project proposes observing and collecting rainfall data from the past ten years to estimate rain for the following day. The work aims to enhance outcomes by finding suitable machine learning systems for foretelling rainfall and comparing the performance of different algorithms like logistic regression, support vector classification, random forest, catboost, and xgboost. Every algorithm will be evaluated based on machine learning parameters such as correctness.