MODELLING MARKET BEHAVIOUR WITH ENSEMBLES AND TECHNICAL INDICATORS
Loading...
Date
Author
Journal Title
Journal ISSN
Volume Title
Publisher
Degree
Master of Science, Information Systems (MScIS)
Discipline
Faculty of Science and Technology
Keywords
Market Event, Ensemble Models, Technical Analysis, Classification
Supervisor(s) and Their Department(s)
Examining Committee Member(s) and Their Department(s)
Degree Grantor
Athabasca University
Abstract
This research investigated the ability of different classification ensemble models to predict the outcome of market events defined by a technical trading system. The ensemble classification models used a diverse set of technical indicators to measure various aspects of market sentiment at the time of market entry as determined by a technical trading system. This research found that various ensemble classification models differ in their ability to classify the nonlinear relationships of market behaviour and are able to perform better than random chance in most cases.