Prediksi Harga Saham dengan Geometric Brownian Motion dan ARIMA - Termodifikasi Kalman Filter

Mustika, Tito Nur (2019) Prediksi Harga Saham dengan Geometric Brownian Motion dan ARIMA - Termodifikasi Kalman Filter. Masters thesis, Institut Teknologi Sepuluh Nopember.

[img]
Preview
Text
06111750012010-Master_Thesis.pdf - Accepted Version

Download (3MB) | Preview

Abstract

Keinginan investor berinvestasi di perdagangan saham, dikarenakan ingin mendapatkan keuntungan dan menghindari kerugian. Perubahan harga saham yang sulit diprediksi, mengakibatkan nilai keuntungan tidak menentu, sehingga dibutuhkan suatu model prediksi. Model Geometric Brownian Motion (GBM) dan model Autoregressive Integrated Moving Average (ARIMA) merupakan dua model pendekatan/aproksimasi yang dipakai untuk memprediksi pergerakan harga saham di masa yang akan datang berdasarkan memori harga saham di masa yang lalu. GBM memprediksi dengan melibatkan perhitungan noise, sedangkan Model ARIMA adalah model yang secara penuh mengabaikan variabel independen dalam membuat prediksi. Pengabaian variabel independen tersebut bertujuan untuk menentukan hubungan statistik variabel dependen dari data masa lalu dengan data masa mendatang. Namun dalam memprediksi, kedua model tersebut masih memiliki error yang cukup besar yang dikarenakan parameter kedua model bersifat konstan. Supaya nilai error tersebut berkurang, maka digunakan metode Kalman-Filter. Berdasarkan hasil yang didapat dalam penelitian ini, hasil nilai MAPE yang didapat dari estimasi GBM-KF dan ARIMA-KF terhadap data faktual termasuk dalam kategori akurasi yang sangat baik. Selanjutnya, perbandingan kedua model yang sudah termodifikasi Kalman-Filter menunjukkan bahwa MAPE untuk model GBM-KF persentase lebih kecil dibanding ARIMA-KF. ================================================================================================ The desire of investors to invest in stock trading, because they want to get profits and avoid losses. Changes in stock prices are difficult to predict, resulting in uncertain profit values, so a prediction model is needed. The Geometric Brownian Motion (GBM) model and the Autoregressive Integrated Moving Average (ARIMA) model are two approach / approximation models used to predict future stock price movements based on past stock price memory. GBM predicts with the calculation of noise, meanwhile The ARIMA model is a model that completely ignores independent variables in making predictions. Ignoring the independent variable aims to determine the statistical relationship of the dependent variable from past data with future data. But in predicting, the two models still have considerable errors caused the parameter of both models are constant. In order for the error value to decrease, the Kalman-Filter method is used. Based on the results obtained in this study, the results of the MAPE values obtained from the GBM-KF and ARIMA-KF estimates of factual data are included in the excellent accuracy category. Furthermore, the comparison of the two modified Kalman-Filter models shows that MAPE for the GBM-KF model is smaller than ARIMA-KF

Item Type: Thesis (Masters)
Additional Information: RTMa 519.287 Mus p-1 2019
Uncontrolled Keywords: Saham, Geometric Brownian Motion (GBM), Autoregressive Integrated Moving Average (ARIMA), GBM-Kalman Filter, ARIMA- Kalman Filter
Subjects: Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry)
Q Science > QA Mathematics > QA402.3 Kalman filtering.
T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44101-(S2) Master Thesis
Depositing User: NUR MUSTIKA TITO
Date Deposited: 14 Apr 2022 06:28
Last Modified: 14 Apr 2022 06:47
URI: https://repository.its.ac.id/id/eprint/69162

Actions (login required)

View Item View Item