EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION

EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION

In this study the ability of support vector regression in forecasting daily rates of the IDX Composite (Indeks Harga Saham Gabungan/IHSG) was investigated. The daily prices of the IDX Composite from 26th of March 2020 to 16th of July 2021 was used. To train the model, 75% of the data was used as tra...

Full description

Saved in:
Bibliographic Details
Main Author: Rafii Manzo Natakusuma, Edgar
Format: Final Project
Language:Indonesia
Online Access:https://digilib.itb.ac.id/gdl/view/60599
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Institut Teknologi Bandung
Language: Indonesia
id id-itb.:60599
spelling id-itb.:605992021-09-18T11:35:08ZEFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION Rafii Manzo Natakusuma, Edgar Indonesia Final Project Hyperparameter, Indeks Harga Saham Sabungan, MAPE, Support Vector Regression, prediction INSTITUT TEKNOLOGI BANDUNG https://digilib.itb.ac.id/gdl/view/60599 In this study the ability of support vector regression in forecasting daily rates of the IDX Composite (Indeks Harga Saham Gabungan/IHSG) was investigated. The daily prices of the IDX Composite from 26th of March 2020 to 16th of July 2021 was used. To train the model, 75% of the data was used as training data, 25% was used as test data, and the last five days were used as hold-out data to see how the three best models would perform on unseen data. To predict, the model was fed the opening, closing, adjusted closing, highest, lowest price of the nth day as input variables and the closing price of the n+5th day as output variables. The RBF kernel SVR model with hyperparameter values of ???? = 100, ???? = 0.1 performed best on test data with a MAPE value of 1,709%. When the three best performing models were tested using the last five days data, the linear SVR with hyperparameter values of ???? = 100, ???? = 0.1 was able to generalize well and produced the least amount of MAPE of 0.364%. text
institution Institut Teknologi Bandung
building Institut Teknologi Bandung Library
continent Asia
country Indonesia
Indonesia
content_provider Institut Teknologi Bandung
collection Digital ITB
language Indonesia
description In this study the ability of support vector regression in forecasting daily rates of the IDX Composite (Indeks Harga Saham Gabungan/IHSG) was investigated. The daily prices of the IDX Composite from 26th of March 2020 to 16th of July 2021 was used. To train the model, 75% of the data was used as training data, 25% was used as test data, and the last five days were used as hold-out data to see how the three best models would perform on unseen data. To predict, the model was fed the opening, closing, adjusted closing, highest, lowest price of the nth day as input variables and the closing price of the n+5th day as output variables. The RBF kernel SVR model with hyperparameter values of ???? = 100, ???? = 0.1 performed best on test data with a MAPE value of 1,709%. When the three best performing models were tested using the last five days data, the linear SVR with hyperparameter values of ???? = 100, ???? = 0.1 was able to generalize well and produced the least amount of MAPE of 0.364%.
format Final Project
author Rafii Manzo Natakusuma, Edgar
spellingShingle Rafii Manzo Natakusuma, Edgar
EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION
author_facet Rafii Manzo Natakusuma, Edgar
author_sort Rafii Manzo Natakusuma, Edgar
title EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION
title_short EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION
title_full EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION
title_fullStr EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION
title_full_unstemmed EFFECT OF KERNEL FUNCTION, C, AND ? HYPERPARAMETER OPTIMIZATION ON PREDICTION ACCURACY ON IHSG USING SUPPORT VECTOR REGRESSION
title_sort effect of kernel function, c, and ? hyperparameter optimization on prediction accuracy on ihsg using support vector regression
url https://digilib.itb.ac.id/gdl/view/60599
_version_ 1822275948349751296

Similar Items