NEW TECHNIQUES
Nonlinear Wave Patterns In The Markets
Singular Spectrum Analysis
Of Price Movement In Forex
by Sergiy Drogobetskii
Looking for new information about price movements? Singular
spectrum analysis may shed some light.
Finding an
analytical method to reduce noise and predict price movement dynamics has
been a popular topic of discussion among traders. In fact, several methods
from the fields of physics and mathematics have already been applied for
this very reason. But how do you extract the necessary information, and
what can you use as the basis for your forecasting model?
SINGULAR SPECTRUM ANALYSIS
Although there are several methods for analyzing time series, they tend
to have limitations. The method I will discuss in this article has some
advantages over other time series analysis techniques: It can tell you
how to extract relevant information from noisy time series and what to
use as a basis for your forecasting model.
Singular spectrum analysis (SSA) is a new analytical method that
has been applied to branches of scientific study such as bioinformatics,
meteorology, astronomy, and pattern recognition. SSA is useful for compressing
information, smoothing of initial data and, in certain cases, predicting
time series data prices. In this article, I will apply SSA to forex market
prices.
Like other financial markets, the forex market is a complex, dynamic
system. Based on my analysis of economic systems, I thought it best to
apply a passive experiment that involves observing the behavior of a system
over time. This resulted in representing the values of observable magnitudes
as a time series. The SSA was designed to provide insight into the dynamics
of the process that generates time series. It is based on the singular
value decomposition (SVD) of a trajectory matrix that is constructed
from the time series of prices.
THE FIRST STEP
Before understanding the principles of SSA, we need some definitions.
I will refer to every price at a fixed point of time as a state of systems.
The set of such states is equidistant in time and forms a one-dimensional
profile of state changes at that particular time.
You are probably aware of the process behind any price movement on a
chart, but you may not be aware of the characteristics of that process.
I will represent this unknown process as a sum of separate components,
which I'll refer to as elementary patterns of behavior (EPB). Each
EPB gives you information about the trend, and the oscillating or noisy
components of the initial time series prices. The singular spectrum analysis
was developed for extracting this information from the initial time series.
...Continued in the October issue of Technical Analysis of
STOCKS & COMMODITIES
Excerpted from an article originally published in the October 2005
issue of Technical Analysis of STOCKS & COMMODITIES magazine. All rights
reserved. © Copyright 2005, Technical Analysis, Inc.
Return to October 2005 Contents
