Pemodelan Curah Hujan Dasarian Periode September Oktober November (SON) di Kabupaten Bantul Menggunakan Hidden Markov Model (HMM)

Putra, Fikri Rizal Dhiya Ul Haq Sandyka (2026) Pemodelan Curah Hujan Dasarian Periode September Oktober November (SON) di Kabupaten Bantul Menggunakan Hidden Markov Model (HMM). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221105-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Curah hujan dasarian pada periode September, Oktober, dan November (SON) memiliki peran penting dalam menggambarkan fase peralihan dari musim kemarau menuju musim hujan. Pada periode tersebut, curah hujan cenderung memiliki variasi yang tinggi sehingga diperlukan pendekatan yang mampu menangkap perubahan kondisi curah hujan dari waktu ke waktu. Penelitian ini bertujuan untuk mengetahui karakteristik curah hujan dasarian serta memperoleh model dan pola probabilitas curah hujan di Kabupaten Bantul menggunakan Hidden Markov Model (HMM). Data yang digunakan merupakan data sekunder curah hujan dasarian periode SON tahun 2004-2023 pada empat lokasi pengamatan, yaitu Piyungan, Dlingo, Gandok, dan Gedongan. Analisis dilakukan melalui statistika deskriptif, estimasi parameter distribusi emisi Gamma, Lognormal, dan Burr XIII menggunakan Maximum Likelihood Estimation (MLE), pemilihan distribusi terbaik berdasarkan Akaike Information Criterion (AIC), estimasi parameter HMM menggunakan algoritma Baum-Welch, serta penentuan urutan hidden state menggunakan algoritma Viterbi. Hasil penelitian menunjukkan bahwa curah hujan dasarian di Kabupaten Bantul didominasi oleh curah hujan rendah dan berpola menceng ke kanan. Gedongan memiliki rata-rata dan variasi curah hujan tertinggi, sedangkan Piyungan memiliki rata-rata terendah. Distribusi Gamma terpilih sebagai distribusi emisi terbaik pada seluruh lokasi pengamatan. Hasil HMM menunjukkan bahwa peluang awal terbesar berada pada state kering dan state kering memiliki persistensi tinggi untuk bertahan pada dasarian berikutnya. Hasil Viterbi menunjukkan dominasi state kering pada awal periode SON, kemudian cenderung bergeser menuju state normal atau basah pada Oktober hingga November. Model HMM mampu menggambarkan pola transisi curah hujan dasarian dan dapat menjadi dasar pendukung dalam memahami dinamika peralihan musim di Kabupaten Bantul. Perbedaan pola antar lokasi menunjukkan adanya variasi spasial dalam respons curah hujan dasarian di Kabupaten Bantul.
======================================================================================================================================
Decadal rainfall during September, October, and November (SON) plays an important role in describing the transition phase from the dry season to the rainy season. During this period, rainfall tends to have high variability, so an approach is needed to capture changes in rainfall conditions over time. This study aims to identify the characteristics of decadal rainfall and obtain a model and probability pattern of rainfall in Bantul Regency using the Hidden Markov Model (HMM). The data used in this study are secondary decadal rainfall data for the SON period from 2004 to 2023 at four observation locations, namely Piyungan, Dlingo, Gandok, and Gedongan. The analysis was conducted through descriptive statistics, parameter estimation of Gamma, Lognormal, and Burr XIII emission distributions using Maximum Likelihood Estimation (MLE), selection of the best distribution based on the Akaike Information Criterion (AIC), HMM parameter estimation using the Baum-Welch algorithm, and determination of the hidden state sequence using the Viterbi algorithm. The results show that decadal rainfall in Bantul Regency is dominated by low rainfall and follows a right-skewed pattern. Gedongan has the highest mean and variability of rainfall, while Piyungan has the lowest mean rainfall. The Gamma distribution was selected as the best emission distribution for all observation locations. The HMM results show that the highest initial probability occurs in the dry state, and the dry state has high persistence to remain in the same state in the following decade. The Viterbi results indicate that the dry state dominates the early SON period and then tends to shift toward normal or wet states in October and November. The HMM is able to describe the transition pattern of decadal rainfall and can serve as supporting information for understanding the dynamics of seasonal transition in Bantul Regency. Differences in patterns among locations indicate spatial variation in the response of decadal rainfall in Bantul Regency.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Algoritma Baum-Welch, Algoritma Viterbi, Curah Hujan Dasarian, Hidden Markov Model, Periode SON, Baum-Welch Algorithm, Decadal Rainfall, SON Period, Viterbi Algorithm
Subjects:	Q Science > QA Mathematics > QA274.2 Stochastic analysis Q Science > QA Mathematics > QA274.7 Markov processes--Mathematical models.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Fikri Rizal Dhiya Ul Haq Sandyka Putra
Date Deposited:	03 Aug 2026 09:34
Last Modified:	03 Aug 2026 09:34
URI:	http://repository.its.ac.id/id/eprint/139894

Actions (login required)

View Item