Stock price prediction using data mining techniques

Biddulph, Max (2015) Stock price prediction using data mining techniques. BSc dissertation, University of Portsmouth.

[img] PDF
Restricted to Registered users only

Download (1105kB)

    Abstract

    To accurately predict the behaviour of the stock market is a highly sought after capability that would bring huge financial gain. It is something that has been attempted many times in a number of different ways over the decades, with different theories and hypotheses as to whether it is even possible. This project reviews existing methods and attempts to implement a data mining approach to stock price prediction using three different data mining algorithms, Linear Regression, M5 Rules and IBK with two different methods.
    Historic data from two companies in the London Stock Exchange (British Petroleum and Tesco) was obtained containing stock market information from the past 15 years. Time series analysis was used with Linear Regression to create models that predict the stock price of a company for a day, a week and a month in the future. The results found that up to a week could be predicted with accuracies of over 99% and a month up to 94%.

    Item Type: Dissertation
    Departments/Research Groups: Faculty of Technology > School of Computing
    Depositing User: Jane Polwin
    Date Deposited: 03 Dec 2015 16:26
    Last Modified: 03 Dec 2015 16:26
    URI: http://eprints.port.ac.uk/id/eprint/19071

    Actions (login required)

    View Item

    Document Downloads

    More statistics for this item...