PENERAPAN METODE RANDOM FOREST DAN XGBOOST UNTUK KLASIFIKASI CUACA PADA PORTABLE AUTOMATIC WEATHER STATION
Farida Nur Azizah, Ir. Unan Yusmaniar Oktiawati, S.T., M.Sc., Ph.D., IPU.
2025 | Tugas Akhir | D4 Teknologi Rekayasa Instrumentasi dan Kontrol
Erratic weather changes impact
various vital sectors, especially agriculture, so an accurate and portable
weather monitoring system such as the Portable Automatic Weather Station is
needed. The device is equipped with temperature and humidity (DHT22), light
intensity (BH1750), wind speed, and rainfall intensity (Rain Gauge). The
classification process requires machine learning algorithms that are not only
accurate but also computationally efficient. This study aims to evaluate and
compare the performance of the Random Forest and XGBoost algorithms in
classifying weather conditions based on Portable Automatic Weather Station
data. Evaluation was carried out on 12,418 data divided into 6 weather classes
by considering aspects of accuracy and computational efficiency. Based on the
results obtained, Random Forest has a slightly superior accuracy of 99.80%
compared to XGBoost with 99.48%. However, XGBoost is superior in terms of
efficiency with a processing time of for 1 minutes, whereas Random Forest takes
as long as 6 minutes. The very small difference in accuracy, which is 0.32%,
with better computing efficiency, makes XGBoost recommended as the right
algorithm to be implemented in the Portable Automatic Weather Station system.
Kata Kunci : Klasifikasi Cuaca, Portable Automatic Weather Station, Random Forest, XGBoost.