Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/125724| Title: | Pemodelan Hidden Markov Kontinu pada Indeks Harga Saham FTSE100 |
| Authors: | Setiawaty, Berlian Ruhiyat Nasrul, Rahmat |
| Issue Date: | 2023 |
| Publisher: | IPB University |
| Abstract: | Harga penutupan indeks pasar saham selalu mengalami perubahan setiap
waktu. Hal ini dipengaruhi salah satunya oleh perubahan kondisi pasar saham.
Kondisi pasar saham yang menyebabkan terjadinya nilai penutupan suatu indeks
saham diasumsikan tidak teramati secara langsung dan membentuk rantai Markov,
sehingga dapat dimodelkan oleh hidden Markov kontinu. Dengan menggunakan
barisan tersembunyi pada hidden Markov diskret, parameter sebaran pada hidden
Markov kontinu dapat diperkirakan. Pada tugas akhir ini akan digunakan model
hidden Markov kontinu untuk memprediksi nilai penutupan harian indeks FTSE
100 periode November 2020 – Desember 2020 dengan data latih pada periode
Januari 2018 – Oktober 2020. Dari hasil perhitungan, nilai penutupan harian
indeks FTSE 100 yang disebabkan oleh tiga state hidden Markov kontinu
memiliki sebaran masing masing state yaitu Weibull(25, 6001, 104),
KhiKuadrat(7284), dan campuran(0.45, 0.55)[seragam (6581, 7877), lognormal (9,
0.01)]. Hasil simulasi dengan Mathematica 12.2 mendapatkan nilai MAPE latih
3.54% dan nilai MAPE uji 5.84%. The closing prices of stock market indices always experience changes over time. This is influenced by changes in market conditions. The market conditions that cause the closing value of a stock index are assumed to be unobservable directly and form a Markov chain, hence, it can be modeled as a continuous hidden Markov. By using the hidden sequence on the discrete hidden Markov, the distribution parameters of continuous hidden Markov can be estimated. In this final project, the hidden Markov model will be used to predict daily closing values of the FTSE 100 index on November 2020 – December 2020, with training data on January 2018 – October 2020. From the calculation results, period, the daily closing values of the FTSE 100 index, which are caused by the three states of the continuous hidden Markov, have the following distributions for each state: Weibull(25, 6001, 104), Chi-Square(7284), and mixture(0.45, 0.55)[uniform(6581, 7877), lognormal(9, 0.01)]. The simulation results using Mathematica 12.2 obtained a training MAPE of 3.54% and a testing MAPE of 5.84%. |
| URI: | http://repository.ipb.ac.id/handle/123456789/125724 |
| Appears in Collections: | UT - Actuaria |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| Dokumen Final Tugas Akhir a.n. Rahmat Nasrul (G94190033).pdf Restricted Access | Fullteks | 11.31 MB | Adobe PDF | View/Open |
| Lampiran Tugas Akhir a.n. Rahmat Nasrul (G94190033).pdf Restricted Access | Lampiran | 3.86 MB | Adobe PDF | View/Open |
| Cover Tugas Akhir a.n. Rahmat Nasrul (G94190033).pdf Restricted Access | Cover | 1.96 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.