Prakiraan Cuaca Jangka Pendek Menggunakan Metode Ridge Regression Dan Principal Component Regression

Fathurrahim, Alif (2024) Prakiraan Cuaca Jangka Pendek Menggunakan Metode Ridge Regression Dan Principal Component Regression. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06211840000099_Undergraduate Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (2MB) | Request a copy

Abstract

Informasi cuaca jangka pendek sangat diperlukan untuk mengetahui kondisi cuaca di esok hari. Informasi cuaca jangka pendek yang penting adalah suhu dan kelembapan. Saat ini pengguna informasi cuaca jangka pendek menuntut kemampuan memperoleh informasi cuaca jangka pendek secara cepat dan akurat. Dengan mempertimbangkan kondisi tersebut, perlu dikembangkan prakiraan cuaca jangka pendek yang cepat dan akurat secara operasional. Salah satu sistem pemodelan cuaca adalah Numerical Weather Prediction (NWP), sesuai dengan tantangan yang terkait dengan metode ilmiah yang memerlukan peramalan cuaca melalui perhitungan fisik sambil mensimulasikan atmosfer yang menantang. Dalam beberapa tahun terakhir, NWP telah banyak digunakan untuk model penelitian dan prakiraan weather research and forecasting (WRF). Salah satu lembaga yang berwenang dan bertanggung jawab secara hukum dalam mengelola data dan informasi cuaca, iklim, dan bencana di Indonesia adalah Badan Meteorologi, Klimatologi, dan Geofisika. Pada tahun 2017, BMKG memiliki 41 radar cuaca yang tersebar di seluruh Indonesia, mulai dari Aceh hingga Papua. Model Output Statistics (MOS) merupakan metode statistik pada tahap pasca pengolahan NWP untuk memperoleh nilai prakiraan parameter cuaca pada suatu titik pengamatan. Masalah dalam pemodelan MOS adalah keadaan data NWP yang berdimensi besar (curse of dimentionality) karena diukur dalam grid yang cukup luas (skala global). Semakin besar dimensi, maka semakin banyak data yang dihasilkan. Oleh karena itu perlu dilakukan pemodelan MOS yang berbasis regresi untuk mengatasi multikolinieritas antar variabel NWP dan memiliki kompleksitas tinggi. MOS dapat di dekati menggunakan metode Ridge Regression dan Principal Component Regression (PCR). Kebaikan model dievaluasi menggunakan Root Mean Square Error (RMSE). Hasil penelitian pada Stasiun Banyuwangi, Stasiun Juanda, dan Stasiun Kalianget menunjukkan bahwa nilai MOS terbaik berdasarkan nilai RMSE yaitu model regresi Ridge.
=================================================================================================================================
Short-term weather information is really necessary to know tomorrow's weather conditions. Temperature and humidity are the important short-term weather information. Currently, users of short-term weather information demand the ability to obtain short-term weather information quickly and accurately. Taking these conditions into account, it is necessary to develop short-term weather forecasts that are rapid and operationally accurate. One of the weather modeling systems is Numerical Weather Prediction (NWP), corresponding to the challenges associated with scientific methods that require forecasting the weather through physical calculations while simulating a challenging atmosphere. In recent years, NWP has been widely used for mesoscale models of weather research and forecasting (WRF). One of the agencies legally authorized and responsible for managing data and information about weather, climate and disasters in Indonesia is the Badan Meteorologi, Klimatologi, dan Geofisika. By 2017, BMKG had 41 weather radars across Indonesia, from Aceh to Papua. Model Output Statistics (MOS) is a statistical method at the post-processing stage of NWP to obtain forecast values for weather parameters at an observation point. The problem in MOS is that the NWP data has large dimensions (curse of dimensionality) because it is measured on a fairly broad grid (global scale). The larger the dimensions, the more data generated. Therefore, it is necessary to carry out regression-based MOS modeling to overcome multicollinearity between NWP variables and which has high complexity. MOS can be approached using the Ridge Regression and Principal Component Regression (PCR) methods. The goodness of the model was evaluated using Root Mean Square Error (RMSE). The results of research at Banyuwangi Station, Juanda Station, and Kalianget Station shows that the best MOS based on the RMSE value is ridge regression model.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Model Output Statistics(MOS), Ridge Regression, Suhu dan Kelembapan, BMKG, Numerical Weather Prediction, Model Output Statistics, Principal Component Regression, Temperature and Humidity
Subjects:	Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA278.5 Principal components analysis. Factor analysis. Correspondence analysis (Statistics)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Alif Fathurrahim
Date Deposited:	30 Aug 2024 08:31
Last Modified:	30 Aug 2024 08:31
URI:	http://repository.its.ac.id/id/eprint/115576

Actions (login required)

View Item