Multi-output machine learning regression for climate prediction: a comparative study of precipitation and temperature forecasting in Jakarta and East Kalimantan, Indonesia

Abstract

The Indonesian government is currently transferring the national capital from Jakarta, on Java Island, to East Kalimantan, on Kalimantan Island. Forecasting climate conditions in both Jakarta and East Kalimantan is crucial for urban planning, disaster risk management, environmental conservation, and economic stability. An accurate climate forecast can guide sustainable development, improve disaster preparedness, and support agriculture, fisheries, and public health. This research focuses on the development of machine learning models for making an accurate climate forecast by proposing some models such as linear regression, polynomial regression, decision tree regressors, K-nearest neighbors regressors, elastic networks, Lasso, and support vector regressors. The predictive models built in this research are multi-output regression for forecasting four climate variables in the next one month as the target or output by using the previous 24 months of data on four climate variables as a feature or input. This research utilized public climate conditions in Jakarta and East Kalimantan obtained from the World Bank Climate Change Knowledge Portal (CCKP). The results show that the linear regression model was the best model for both forecasting scenarios in Jakarta and East Kalimantan.