Discovering associative price patterns in stock data

Pao Yue-kong Library Electronic Theses Database

Discovering associative price patterns in stock data

 

Author: Hu, Lun
Title: Discovering associative price patterns in stock data
Degree: M.Sc.
Year: 2008
Subject: Hong Kong Polytechnic University -- Dissertations.
Stock price forecasting.
Investment analysis.
Department: Dept. of Computing
Pages: x, 100 p. : ill. ; 30 cm.
Language: English
InnoPac Record: http://library.polyu.edu.hk/record=b2240170
URI: http://theses.lib.polyu.edu.hk/handle/200/3704
Abstract: Accompanying with the development of economy, the stock trade has already being an important investment tool for people to deposit or increase their wealth. However, when referring to the analysis of stocks, most people determine the actions (i.e., buy or sell) according to their subjective experience as well as the performance of companies. Though the accuracy resulting from those common methods may be encouraging for some time, they ignore the similarity and regularity of data, thus being lack of both objectivity and rationality. Hence, more and more researchers focus their attention to develop objective and computational methods achieving the goal of analyzing stocks with a good performance, and data mining is a hot general approach in this field. Since data mining intends to tackle this issue in the way of finding similarity hidden in data, the appearances of data used as input include numerical or categorical. However, in order to represent different shapes of stock price movement, we introduce the conception of price pattern, consisting of a sequence of categorical values, to represent those shapes. This dissertation proposes a statistical computational method to determine whether or not associative price patterns exist in financial time series such as in stock price series. The problem of finding patterns in financial time series has been tackled by systematical observations, statistical analysis or by the use of artificial intelligence techniques. However, the emphasis is more on finding pattern in data rather than finding associations between patterns. Given that patterns in time series may overlap or be scattered within a series, the discovering of pattern association has not receive much attention. The dissertation proposes a technique to find elementary patterns in financial time series. Based on it, a sequence of patterns, (yi)n, i=0 can then be constructed for the original time series. Using the idea of residual analysis, we show how we can conclude that an associative relationship exists between two patterns yn and yn-1.

Files in this item

Files Size Format
b22401702.pdf 4.357Mb PDF
Copyright Undertaking
As a bona fide Library user, I declare that:
  1. I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
  2. I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
  3. I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.

     

Quick Search

Browse

More Information