Analysis of Temperature Prediction Using Random Forest and Facebook Prophet Algorithms

Forecasting temperature daily has been a great challenge that the meteorological department faces today. For Kerala, the year 2019 has one of the warmest summers on record. Increase in temperature has adverse effects on health and agriculture fields. Accurate prediction of daily temperature enables the Government and people to take proper precautionary steps. This paper focuses on analysing two algorithms-Random Forest and Facebook Prophet- for temperature prediction. Their performance-based on five different stations in Kerala, India, are compared based on Accuracy and Mean Absolute Error. Both the models gave comparable results, but Random Forest gave better accuracy and Mean absolute error. The forecasts of Random Forest is more consistent than Facebook Prophet.