PERFORMANCE ANALYSIS OF LIGHTGBM MODEL IN PREDICTING RAIN INTENSITY IN THE KUALANAMU CLASS 1 METEOROLOGICAL STATION AREA

Abstract

The intensity of rain or rainfall that occurs is influenced by various weather parameters and it plays a big role for the community. Therefore, information related to rain intensity is very important, there is a need for the availability of information related to this. This study aims to analyze the performance of Machine Learning using the Light Gradient Boosting Machine model in predicting the intensity of rainfall in the Kualanamu Meteorological Station area during the 2018-2022 time span. Historical data collection is done through synoptic data collection that has been issued by Kualanamu class 1 Meteorological Station. Several matrix evaluations are used in the form of Accuracy, AUC (Area Under the Curve), Recall, Precision, and F1 Score. The matrix evaluation is able to produce detailed evaluation calculations and is able to measure how well the model works. Then the average value of the matrix evaluation is 0.7251 for accuracy, 0.8122 for AUC, 0.7251 for Recall, 0.7236 for Precision and 0.7231 for F1 Score. Based on the results obtained, the Light Gradient Boosting Machine model is able to provide good rain intensity prediction results but there is a need for further analysis in the model development stage, focusing on reducing the error rate and increasing prediction accuracy so as to make a significant contribution to the planning and decision-making process related to weather conditions.