Hedianti, Elsa Siffana (2019) Peramalan Harga Saham Dengan Menggunakan Metode Support Vector Regression (SVR). Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
05211540000014-Undergraduate_Theses.pdf Download (2MB) | Preview |
Abstract
Harga saham sangat berpengaruh dalam perkembangan ekonomi. Permasalahan ketidakpastian harga saham memiliki risiko yang besar bagi para investor dan pelaku bisnis terkait. Hal tersebut membuat peramalan harga saham yang akurat dapat membantu dalam pengambilan keputusan finansial. Data harga saham merupakan data finansial yang kompleks, memiliki banyak noise, bersifat non – linear dan tidak stasioner, oleh karenanya diperlukan model yang mampu meramalkan dengan ciri – ciri data seperti data finansial tersebut.
Support Vector Mechine Regression (SVR) adalah adaptasi dari machine learning berdasar klasifikasi model regresi dari Support Vector Machine (SVM). SVR merupakan metode yang dapat menyelesaikan permasalahan estimasi non-linear sehingga bisa digunakan untuk meramalkan harga saham. Peramalan harga saham dengan menggunakan model SVR akan dibantu dengan optimasi parameter yaitu Fruit Fly Optimization Algorithm (FOA). Adapun variabel terikat yang digunakan dalam penelitian ini adalah data close price harga saham harian dan variabel terbuka adalah data open price, high price, low price, volume harga saham harian, serta data exchange rate IDR-USD.
Model SVR dengan bantuan FOA untuk pencarian parameter optimal dapat digunakan untuk meramalkan harga saham harian. Hasil analisa menunjukkan bahwa model SVR dengan parameter C adalah 9139.009142607989, ε adalah 1.0219421008384209, dan γ adalah 381.67717950346355 merupakan model SVR yang terbaik untuk melakukan peramalan harga saham harian. Hal ini ditunjukkan dari nilai MAPE yang didapatkan adalah 0.020666150003759407%.
Hasil peramalan masa mendatang dengan menggunakan metode SVR menghasilkan nilai akurasi yang lebih baik dibandingkan dengan moving average (MA) yaitu dengan nilai MAPE peramalan model SVR adalah 0,0023% lebih baik dibandingkan nilai MAPE peramalan model MA yaitu 2,3%.
================================================================================================
Stock prices are very influential in economic development. The issue of uncertainty in stock prices has a big risk for investors and related business people. This makes accurate stock price forecasting can help in financial decision making. Stock price data is complex financial data, has a lot of noise, is non - linear and not stationary, therefore a model that is capable to forecast with data characteristics like those financial data is needed.
Support Vector Mechine Regression (SVR) is an adaptation of machine learning based on regression model classification of Support Vector Machine (SVM). SVR is a method that can solve non-linear estimation problems so that it can be used to forecast stock prices. Forecasting stock prices using the SVR model will be assisted by parameter optimization, that is Fruit Fly Optimization Algorithm (FOA). The dependent variable used in this study is close price data of daily stock prices and the open variables are data open price, high price, low price, volume of daily stock prices, and IDR-USD exchange rate data.
The SVR model with FOA assistance for optimal tuning parameter can be used to forecast daily stock prices. The analysis result shows that the SVR model with C parameter value 9139.009142607989, ε value 1.0219421008384209, and γ value 381.67717950346355 is the best SVR model for forecasting daily stock prices. This is indicated by the MAPE value obtained, that is 0.020666150003759407%.
The result of future forecasting by using SVR method give a better accuracy compared to Moving Average (MA) method with MAPE value of SVR model was 0.0023% better than MAPE value of MA model, that was 2,3%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Additional Information: | RSSI 519.535 Hed p-1 2019 |
| Uncontrolled Keywords: | Harga Saham, Forecasting, Support Vector Regression, Fruit Fly Optimization Algorithm |
| Subjects: | H Social Sciences > H Social Sciences (General) > H61.4 Forecasting in the social sciences H Social Sciences > HA Statistics > HA31.3 Regression. Correlation Q Science > QA Mathematics > QA278.2 Regression Analysis. Logistic regression Q Science > QA Mathematics > QA280 Box-Jenkins forecasting Q Science > QA Mathematics > QA353.K47 Kernel functions (analysis) T Technology > T Technology (General) > T57.6 Operations research--Mathematics. Goal programming |
| Divisions: | Faculty of Information and Communication Technology > Information Systems > 57201-(S1) Undergraduate Thesis |
| Depositing User: | Elsa Siffana |
| Date Deposited: | 05 Jul 2021 05:13 |
| Last Modified: | 05 Jul 2021 05:13 |
| URI: | http://repository.its.ac.id/id/eprint/60693 |
Actions (login required)
 |
View Item |