Prediksi Imbal Hasil Obligasi Menggunakan Model Tiga Faktor Berbasis Autoencoder dan Neural Network Augmented State-Space

Taufiq, Almira Salsabila (2023) Prediksi Imbal Hasil Obligasi Menggunakan Model Tiga Faktor Berbasis Autoencoder dan Neural Network Augmented State-Space. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06111940000114-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (19MB) | Request a copy

Abstract

Investasi merupakan salah satu faktor yang dapat meningkatkan produktivitas dan efisiensi perekonomian negara. Obligasi sebagai sarana investasi tentu membutuhkan imbal hasil dari instrumen tersebut agar para investor modal tertarik untuk menanamkan sejumlah dana mereka di dalamnya. Salah satu cara yang digunakan untuk mengetahui hubungan antara imbal hasil yang diperoleh dengan waktu jatuh tempo untuk suatu obligasi pada waktu tertentu adalah melalui kurva imbal hasil. Karenanya, keakuratan dalam menganalisis bentuk kurva imbal hasil penting untuk memprediksi fluktuasi kurva tersebut. Penelitian ini menggunakan model berbasis jaringan saraf dengan perluasan metode tiga faktor Nelson-Siegel dalam meningkatkan kemampuan prediktif untuk mendapatkan muatan faktor yang serupa dengan model kurva imbal hasil. Diberikan struktur model Gaussian linear state-space model pada model faktor kurva hasil berbasis jaringan saraf dalam model Neural Network Augmented State-Space (NNSS) yang menghasilkan parameter fungsi pemuatan faktor, yang dapat diestimasi bersama dengan parameter transisi. Diperoleh hasil perhitungan evaluasi menggunakan MAE dimana model berbasis jaringan saraf memiliki nilai akurasi yang lebih baik. Diketahui pula, secara umum, nilai model dengan empat laten memiliki hasil yang lebih baik dan lebih stabil dibandingkan model berbasis lima laten. Berdasarkan hasil prediksi imbal hasil, didapatkan bahwa untuk jatuh tempo dengan jangka waktu yang lebih panjang, nilai imbal hasil yang diterima investor akan semakin besar yang membuat obligasi dapat dikatakan undervalued.
==================================================================================================================================
Investment is one of the factors that can increase the productivity and efficiency of a country's economy. Bonds as a means of investment certainly require a return on the instrument so that capital investors are interested in investing some of their funds in it. One of the ways used to determine the relationship between the yield obtained and the maturity time for a bond at a certain time is through the yield curve. Therefore, accuracy in analyzing the shape of the yield curve is important to predict fluctuations in the curve. This study uses a neural network-based model with an extension of the Nelson-Siegel three-factor method to improve the predictive ability to obtain a factor load similar to the yield curve model. Given a Gaussian linear state-space model structure on the neural network-based yield curve factor model in the Neural Network Augmented State-Space (NNSS) model that generates factor loading function parameters, which can be estimated along with transition parameters. The results of the evaluation calculation using MAE where the neural network-based model has a better accuracy value are obtained. It is also known that, in general, the value of the model with four latents has better and more stable results than the model based on five latents. Based on the yield prediction results, it is found that for maturities with a longer period of time, the value of the yield received by investors will be greater, which makes the bonds undervalued.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kurva imbal hasil, model Nelson-Siegel, neural network; Yield curve, nelson-siegel model, neural network
Subjects:	H Social Sciences > HG Finance Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) R Medicine > R Medicine (General) > R858 Deep Learning
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Almira Salsabila Taufiq
Date Deposited:	25 Sep 2023 01:18
Last Modified:	25 Sep 2023 01:18
URI:	http://repository.its.ac.id/id/eprint/102796

Actions (login required)

View Item