Peramalan Total Nominal Klaim Pada Program Jaminan Hari Tua (JHT) Menggunakan Metode Hybrid ARIMA-Long Short Term Memory (ARIMA-LSTM)

Margiansyah, Syafarizal Irgi (2024) Peramalan Total Nominal Klaim Pada Program Jaminan Hari Tua (JHT) Menggunakan Metode Hybrid ARIMA-Long Short Term Memory (ARIMA-LSTM). Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 5006201072-Undergraduate_thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (3MB) | Request a copy

Abstract

Risiko dalam berbagai sektor semakin meningkat di era ini, mendorong perlunya manajemen risiko yang bijaksana sebagai langkah preventif untuk melindungi individu, organisasi, perusahaan, dan lembaga dari potensi kerugian. Salah satu sektor yang sangat rentan terhadap risiko keuangan adalah industri asuransi, yang baru-baru ini disoroti oleh serangkaian kebangkrutan perusahaan asuransi ternama di Indonesia. Dalam konteks ini, Badan Penyelenggara Jaminan Sosial (BPJS) Ketenagakerjaan memegang peran sentral dalam mengelola risiko terkait jaminan sosial, termasuk program Jaminan Hari Tua (JHT), yang memberikan perlindungan finansial bagi pekerja setelah masa pensiun. Penelitian ini bertujuan untuk meramalkan data total nominal klaim JHT menggunakan metode hybrid ARIMA-LSTM, dengan data periode Januari 2018 hingga Februari 2024. Metode hybrid ARIMA-LSTM dipilih karena efektif dalam meramalkan data time series dalam berbagai bidang, seperti yang telah ditunjukkan dalam penelitian terdahulu dalam berbagai bidang, termasuk prediksi ekspor dan kasus COVID-19. Melalui penelitian ini, didapatkan hasil model penelitian terbaik yaitu Hybrid ARIMA (1,1,2) dengan LSTM (2-2-1) dengan lag signifikan
y_(t-3) dan y_(t-12). Model tersebut mendapatkan nilai evaluasi 16,9205%. Model Hybrid ARIMA (1,1,2) dengan LSTM (2-2-1) menurunkan nilai error sebesar 1,3139% dari model ARIMA. Menggunakan model terbaik tersebut, didapatkan hasil peramalan total nominal klaim (JHT) dengan range nominal sejumlah 44 miliar hingga mendekati 47 miliar Rupiah. Dengan rata-rata 46 miliar untuk prediksi satu tahun kedepan. Hasil peramalan ini dapat digunakan sebagai dasar bagi BPJS Ketenagakerjaan untuk menyusun strategi keuangan yang lebih efektif dan mengelola likuiditas dengan lebih baik.
=================================================================================
Risks in various sectors are increasing in this era, encouraging the need for wise risk management as a preventive measure to protect individuals, organizations, companies and institutions from potential losses. One sector that is particularly vulnerable to financial risk is the insurance industry, which has recently been highlighted by a series of bankruptcies of well-known insurance companies in Indonesia. In this context, the Social Security Administering Agency (BPJS) for Employment plays a central role in managing risks related to social security, including the Old Age Security (JHT) program, which provides financial protection for workers after retirement. This research aims to predict the total nominal data for JHT claims using the method hybrid ARIMA-LSTM, with data for the period January 2018 to February 2024. The ARIMA-LSTM hybrid method was chosen because it is effective in predicting time series data in various fields, as has been shown in previous research in various fields, including predicting exports and COVID-19 cases. Through this research, the best research model results were obtained, namely Hybrid ARIMA (1,1,2) with LSTM (2-2-1) with significant lags y_(t-3) and y_(t-12).. The model received an evaluation score of 16.9205%. Model Hybrid ARIMA (1,1,2) with LSTM (2-2-1) lowers the value error of 1.3139% from the ARIMA model. Using the best model, the total nominal claim (JHT) forecasting results were obtained with range nominal amount of 44 billion to close to 47 billion Rupiah. With an average of 46 billion for predictions for the next year. The results of this forecast can be used as a basis for BPJS Employment to develop more effective financial strategies and manage liquidity better.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Hybrid ARIMA-LSTM, Jaminan Hari Tua (JHT), Kerugian, Manajemen risiko, dan Peramalan Hybrid ARIMA-LSTM, Losses Old Age Security (JHT), Risk Management, and Forcasting
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Actuaria > 94203-(S1) Undergraduate Thesis
Depositing User:	Syafarizal Irgi Margiansyah
Date Deposited:	01 Aug 2024 02:24
Last Modified:	01 Aug 2024 02:24
URI:	http://repository.its.ac.id/id/eprint/110267

Actions (login required)

View Item