Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/138465
Title: Machine learning applications in Singapore Stock Exchange
Authors: Liew, Jia Hao
Ho, Stefanie Jia Jia
Ko, Yu Zhuang
Keywords: Social sciences::Economic development
Issue Date: 2020
Publisher: Nanyang Technological University
Project: HE_1AY1920_20
Abstract: There is a lack of research on the application of machine learning for stock return prediction in small stock exchanges like the Singapore Exchange (SGX), despite the existence of substantial evidence on the success of machine learning application onto stock returns in global markets such as in the US. Our first objective is to investigate the effectiveness of machine learning on stock returns prediction for SGX by utilising both linear and nonlinear machine learning models. Our second objective is to perform a deeper analysis into the industry level to understand the effectiveness of each machine learning technique on the various industries and lastly, to find out which predictive variables are the most significant in Singapore’s context. Our study utilises a total of seven machine learning models, from the simplest linear regression (OLS) to nonlinear models such as boosted regression tree and random forest. Our findings suggest that firstly, machine learning is effective for SGX, showing improvement in out-of-sample R2 values in comparison to the base OLS model. Also, nonlinear models greatly outperform the linear models. The three most important variables across the models are the six-month price momentum, turnover rate of shares, and illiquidity. On the industry level, performance of machine learning varies for different industries. Similar to a stock return prediction study on the US’s stock market, momentum has shown to be an important factor for linear models, for both large and small markets. Also, nonlinear models for both studies uncover non-momentum variables as the most important; dividend-price ratio for the US, and valuation ratios for a small stock exchange like SGX.
URI: https://hdl.handle.net/10356/138465
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SSS Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
Machine Learning Applications in Singapore Stock Exchange.pdf
  Restricted Access
491.65 kBAdobe PDFView/Open

Page view(s) 50

316
Updated on Jun 14, 2021

Download(s) 50

75
Updated on Jun 14, 2021

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.