Financial time series predicting using machine learning algorithms

Financial time series prediction is a challenging task due to the fluctuation of trading or economic exchange that is difficult to predict. Researchers from different fields have been attracted to perform several techniques for identifying reliability of the financial time series prediction. Finding...

Full description

Bibliographic Details
Main Author: Tiong, Leslie Ching Ow *
Format: Thesis
Published: 2013
Subjects:
Online Access:http://eprints.sunway.edu.my/234/
_version_ 1848801778882052096
author Tiong, Leslie Ching Ow *
author_facet Tiong, Leslie Ching Ow *
author_sort Tiong, Leslie Ching Ow *
building SU Institutional Repository
collection Online Access
description Financial time series prediction is a challenging task due to the fluctuation of trading or economic exchange that is difficult to predict. Researchers from different fields have been attracted to perform several techniques for identifying reliability of the financial time series prediction. Finding of research papers, the financial trend patterns repeat itself in the history. Thus, this research motivates and aims to investigate the repeat behaviour and pattern of trends from the historical financial time series data, and utilise the strength of machine learning techniques to develop a promising financial time series predictor engine. In this research, two frameworks are proposed for financial time series prediction. In the first proposed framework, candlestick pattern is utilised as technical analysis method to identify the financial trends. Thereafter, Artificial Neural Network (ANN) and Support Vector Machine (SVM) algorithms are implemented separately to train with the trend patterns for predicting the movement direction of financial trends. In the second proposed framework, Linear Regression Line (LRL) is utilised to identify the trend patterns from historical financial time series, which is supported by ANN and SVM for classification process separately. Subsequently, Dynamic Time Warping (DTW) algorithm is utilised through brute force to predict the trend movement. The experimental results showed that the second proposed model is consistent with the hypothesis, which provides better accuracy of prediction. Therefore, the findings of this research help in improving the accuracy of prediction model.
first_indexed 2025-11-14T21:12:52Z
format Thesis
id sunway-234
institution Sunway University
institution_category Local University
last_indexed 2025-11-14T21:12:52Z
publishDate 2013
recordtype eprints
repository_type Digital Repository
spelling sunway-2342020-09-22T02:13:21Z http://eprints.sunway.edu.my/234/ Financial time series predicting using machine learning algorithms Tiong, Leslie Ching Ow * HG Finance QA76 Computer software Financial time series prediction is a challenging task due to the fluctuation of trading or economic exchange that is difficult to predict. Researchers from different fields have been attracted to perform several techniques for identifying reliability of the financial time series prediction. Finding of research papers, the financial trend patterns repeat itself in the history. Thus, this research motivates and aims to investigate the repeat behaviour and pattern of trends from the historical financial time series data, and utilise the strength of machine learning techniques to develop a promising financial time series predictor engine. In this research, two frameworks are proposed for financial time series prediction. In the first proposed framework, candlestick pattern is utilised as technical analysis method to identify the financial trends. Thereafter, Artificial Neural Network (ANN) and Support Vector Machine (SVM) algorithms are implemented separately to train with the trend patterns for predicting the movement direction of financial trends. In the second proposed framework, Linear Regression Line (LRL) is utilised to identify the trend patterns from historical financial time series, which is supported by ANN and SVM for classification process separately. Subsequently, Dynamic Time Warping (DTW) algorithm is utilised through brute force to predict the trend movement. The experimental results showed that the second proposed model is consistent with the hypothesis, which provides better accuracy of prediction. Therefore, the findings of this research help in improving the accuracy of prediction model. 2013-09 Thesis NonPeerReviewed Tiong, Leslie Ching Ow * (2013) Financial time series predicting using machine learning algorithms. Masters thesis, Sunway University.
spellingShingle HG Finance
QA76 Computer software
Tiong, Leslie Ching Ow *
Financial time series predicting using machine learning algorithms
title Financial time series predicting using machine learning algorithms
title_full Financial time series predicting using machine learning algorithms
title_fullStr Financial time series predicting using machine learning algorithms
title_full_unstemmed Financial time series predicting using machine learning algorithms
title_short Financial time series predicting using machine learning algorithms
title_sort financial time series predicting using machine learning algorithms
topic HG Finance
QA76 Computer software
url http://eprints.sunway.edu.my/234/