Prediction of Drought - A Machine Learning Approach using Time Series Data

In order to forecast drought in a certain area, this research employed three different machine learning models, namely ARIMA (Auto Regressive Integrated Moving Average Model), VAR (Vector Auto Regressive model), and Prophet. These models are used to forecast upcoming drought occurrences and are trained using historical data on a variety of meteorological factors, including temperature, precipitation, humidity, and wind speed. Different statistical metrics, such as mean absolute error, root mean squared error, and coefficient of determination, are used to assess each model's performance. The findings indicate that all three models are capable of accurately and reliably forecasting drought episodes, with Prophet surpassing the other models. According to the findings of the study, machine learning models can be useful tools for forecasting drought episodes in a particular region, giving decision-makers a head start on taking the necessary actions to lessen the effects of drought on agriculture, water resources, and society. Among the three time series models experimented, Prophet model exhibited the accuracy of 92.01% which is the best when compared with the other two models. The choice of selecting a model for predicting the drought depends on the type of data being used for experimentation.