Prakiraan Suhu Dan Kelembapan Jangka Pendek Menggunakan Metode Partial Least Square Regression (Plsr) Dan Principal Component Regression (Pcr)

Sudiro, Muhammad Arif (2022) Prakiraan Suhu Dan Kelembapan Jangka Pendek Menggunakan Metode Partial Least Square Regression (Plsr) Dan Principal Component Regression (Pcr). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211840000077-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (3MB)

Abstract

Cuaca memiliki pengaruh yang sangat vital terhadap kelangsungan dan kelancaran aktivitas masyarakat di seluruh belahan dunia, tidak terkecuali Indonesia. Informasi tentang prakiraan cuaca yang cepat dan akurat menjadi suatu hal yang penting karena cuaca menjadi bagian yang tak terpisahkan dari aktifitas manusia dan mempengaruhi berbagai aktivitas masyarakat, salah satunya bidang transportasi. Badan Meteorologi, Klimatologi, dan Geofisika (BMKG) sebagai lembaga pemerintah yang berperan dalam prediksi cuaca mengeluatkan beberapa prediksi cuaca berdasarkan data output dari Numerical Weather Prediction (NWP). Namun NWP masih menghasilkan ramalan yang bias. Oleh karena itu, dilakukan pemrosesan secara statistik pada data output NWP menggunakan model output statistics (MOS), yaitu pemodelan hubungan antara hasil observasi cuaca dan output NWP berbasis regresi. MOS dapat didekati dengan metode Partial Least Square Regression (PLSR) dan Principal Component Regression (PCR). Dari kedua model tersebut dilakukan kalibrasi ensemble dengan menggunakan ensemble model output statistics (EMOS). Kebaikan model untuk mendapatkan model terbaik dievaluasi menggunakan kriteria root mean square error (RMSE) dan Continuous Ranked Probability Score (CRPS). Hasil penelitian pada kedua stasiun pengamatan menunjukkan bahwa model MOS terbaik berdasarkan nilai RMSE yaitu model regresi PLSR. Untuk prediksi terkalibrasi berdasarkan nilai CRPS yang diperoleh, model EMOS lebih terkalibrasi dibandingkan dengan raw ensemble dalam memprediksi temperatur dan kelembaban.
==============================================================================================================================
Weather has a very vital influence on the continuity and smoothness of community activities in all parts of the world, Indonesia is no exception. Information about fast and accurate weather forecasts is important because weather is an inseparable part of human activities and affects various community activities, one of which is the transportation sector. The Meteorology, Climatology and Geophysics Agency (BMKG) as a government agency that plays a role in weather prediction issues several weather predictions based on output data from Numerical Weather Prediction (NWP). But the NWP still produces biased forecasts. Therefore, statistical processing is carried out on the NWP output data using the output statistics (MOS) model, namely modeling the relationship between the results of weather observations and regression-based NWP output. MOS can be approached using Partial Least Square Regression (PLSR) and Principal Component Regression (PCR) methods. From the two models, ensemble calibration was carried out using ensemble model output statistics (EMOS). The goodness of the model to get the best model was evaluated using the criteria of root mean square error (RMSE) and Continuous Ranked Probability Score (CRPS). The results of the research at the two observation stations showed that the best MOS model based on the RMSE value was the PLSR regression model. For calibrated predictions based on the obtained CRPS values, the EMOS model is more calibrated than the raw ensemble in predicting temperature and humidity.

Item Type:	Thesis (Other)
Additional Information:	RSSt 519.536 Sud p-1 2022
Uncontrolled Keywords:	BMKG, Model Output Statistics (MOS), Numerical Weather Prediction (NWP), Partial Least Square Regression (PLSR), Principal Component Regression (PCR), Suhu dan kelembapan. BMKG, Model Output Statistics (MOS), Numerical Weather Prediction (NWP), Partial Least Square Regression (PLSR), Principal Component Regression (PCR), temperature and humidity
Subjects:	H Social Sciences > HA Statistics
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Mr. Marsudiyana -
Date Deposited:	11 Jun 2026 03:02
Last Modified:	11 Jun 2026 03:02
URI:	http://repository.its.ac.id/id/eprint/133722

Actions (login required)

View Item