Publication detail

Automatic head and shoulders pattern recognition using nonlinear statistical modelling

Authors: Heckenbergerová Jana | Marek Jaroslav | Mejznar Jakub

Year: 2013

Type of publication: ostatní - článek ve sborníku

Name of source: APLIMAT 2013, 12th Conference on Applied Mathematics, Book of Abstracts

Publisher name: Slovenská technická univezita v Bratislave

Place: Bratislava

Page from-to: 58

Titles:

Language	Name	Abstract	Keywords
eng	Automatic head and shoulders pattern recognition using nonlinear statistical modelling	Chart Pattern recognition is one of the keystone of Technical Analysis, discipline that concentrates all methods of market development evaluation. Graphical formations such as Head and Shoulders, Flag or Triangle are usually detected only visually by experts long after their appearance. The aim of presented contribution is introduction of new automatic algorithm for Head and Shoulders Pattern (HaSP) detection. The algorithm is based on nonlinear function approximation and statistical estimation of unknown parameters. Whole procedure is tested in Case Study consisting long time series of Dow Jones average indexes. In the end, the successfulness of different market strategies is tested by statistical analysis of percentage profit. Results show that pattern appearance does not have significant effect on prices development.	Technical analysis, Chart patterns, Market Strategies, Nonlinear approximation, Regression model, Linearization, Index of determination.

Search