DSpace Repository

Predicting Short-Term Stock Movements Using Limit Order Book Dynamics and Support Vector Machines in the Stock Exchange of Thailand

Show simple item record

dc.contributor.advisor Guha, Sumanta
dc.contributor.author Manandhar, Sujal
dc.contributor.other Esichaikul, Vatcharaporn
dc.contributor.other Yingsaeree, Chaiyakorn
dc.date.accessioned 2015-05-12T07:05:29Z
dc.date.available 2015-05-12T07:05:29Z
dc.date.issued 2015-05
dc.identifier.uri http://www.cs.ait.ac.th/xmlui/handle/123456789/743
dc.description.abstract Today, market information is readily available to the people and their trading behaviour is significantly affected by the current market conditions. Incorporating the market conditions into the forecasting model will improve the predictive performance and one way to achieve this is by modeling the limit order book dynamics. The aims of this research were to develop a predictive model that using the limit order book dynamics, predicts the direction of the stock movement and to determine the features that significantly affect the predictive performance of the model. We developed models based on a novel application of support vector machines (SVMs). The historical limit order book data of Stock Exchange of Thailand (SET) is mined to obtain informative value which is used by the model for the forecasting purpose. Different feature sets were used. Results showed that the data in the lower part of the order book, mostly volumes exhibit predictive information. The overall evaluation parameter of the models are high. However, it was found that this is mainly from the contribution of the stationary market cases. So even though the model can be used for our intended purpose, it is suggested at the end, considering the precision of results, that the model be used to determine the state of the market i.e. whether it remains stationary or it moves but not the direction of the movement. en_US
dc.language.iso en en_US
dc.subject predictive model en_US
dc.subject limit order book en_US
dc.subject stocks en_US
dc.subject SVM en_US
dc.subject SET en_US
dc.subject information gain en_US
dc.title Predicting Short-Term Stock Movements Using Limit Order Book Dynamics and Support Vector Machines in the Stock Exchange of Thailand en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account