Professional Documents
Culture Documents
Wavelet Multi-Resolution Analysis of High Frequency FX Rates
Wavelet Multi-Resolution Analysis of High Frequency FX Rates
Wavelet Multi-resolution
Analysis of High
Frequency FX Rates
Department of Computing
University of Surrey, Guildford, UK
August 27, 2004
Talk Outline
400
350
300
250
200
83
84
85
86
Year
87
TS Data
Irregular
Cyclical
Trend Component
Indicates the very long-term behavior of the
time series
Typically as a straight line or an exponential
curve
This is useful in seeing the overall picture
Cyclical Component
Peak
Contraction
Trough
Expansion
Time
Seasonal Component
Time
Irregular Component
Random, unsystematic, residual
fluctuations
Due to random variation or unforeseen
events
Short duration and non-repeating
A forecast, even in the best situation, can be
no closer (on average) than the typical size
of the irregular variation
IBM Prices
505
465
Nonstationary
425
385
345
305
1
26
51
76
101
126
151
176
201
226
251
276
301
326
351
0.1
IBM Volatility
0.08
0.06
Time Varying
Volatility
0.04
0.02
0
1
26
51
76
101
126
151
176
201
226
251
276
301
326
351
Signal
Wavelet
C = C1
Signal
Wavelet
C = C2
Wavelet
C = C3
w* xt wi xt i
i
Xt Original Signal
W High or low pass filters
Xt
H* (f)
D1
G (f)
A1
G* (f)
A1
Frequency
H (f)
Xt
A2
L
Level 1
H
D2
L
D1
Xt = A1 + D1
Level 2
Xt = A2 + D1+ D2
Level 3
A3
D3
Xt = A3 + D1+ D2 + D3
Level N
Xt = AN + D1+ D2 + DN
3.5E+05
4.8E+03
2.5E+05
4.6E+03
4.4E+03
1.5E+05
2.0E+05
1.0E+05
4.2E+03
4.0E+03
5.0E+04
3.8E+03
0.0E+00
3.0E+06
-1.0E+02
-1.5E+02
1.5E+06
1.0E+06
5.0E+05
0.0E+00
3.0E+06
2.5E+06
1.5E+02
2.0E+06
5.0E+01
1.5E+06
-5.0E+01
5.0E+05
1.0E+06
-1.5E+02
8.9E+01
FFT (2)
Level - 2
2.5E+02
2.0E+06
FFT (3)
-5.0E+01
2.5E+06
0.0E+00
3.5E+05
3.0E+05
3.9E+01
-1.1E+01
-6.1E+01
-1.1E+02
2.5E+05
2.0E+05
1.5E+05
1.0E+05
5.0E+04
0.0E+00
FFT (1)
Level - 3
0.0E+00
3.0E+05
FFT (FTSE)
5.2E+03
5.0E+03
5.0E+01
Level - 1
Wavelet Decompositions
FTSE 100
Analyzing High-frequency
Financial Data: Our Approach
Tick Data
Preprocessing
Transformation
Knowledge
Discovery
Forecast
Data
Compression
Multiscale
Analysis
Summarization
Prediction
Aggregate the
movement in the
dataset over a
certain
period of time
Describe market
dynamics at
different scales
(time horizons)
with
chief features
Use the
extracted
chief features
to predict
Cycle
Trend
Turning Points
Variance Change
Analyzing High-frequency
Financial Data: Our Approach
Generalized Algorithm: Summarization
I.
Compress the tick data to get Open (O), High (H), Low (L) and Close (C)
value for a given compression period (for example, one minute or five
minutes).
II.
III.
III-1.
III-2.
III-3.
IV.
Locate a single variance change in the series by using the NCSS index on
C.
V.
Generate a graphical and verbal summary for results of Steps III-1 to III-3
and IV.
Analyzing High-frequency
Financial Data: Our Approach
Generalized Algorithm: Prediction
I.
Summarize the tick data using the time series summarization algorithm.
II.
III.
IV.
Add the results of Steps II and III to get an aggregate N-step ahead forecast,
forecast
Forecast = DS, forecast + AN, forecast.
Analyzing High-frequency
Financial Data: Our Approach
A prototype system has been implemented that automatically
extracts chief features from a time series and give a prediction
based on the extracted features, namely trend and seasonality
Raw Signal
Volatility
NCSS
Detect
Variance
Change
Prediction
Statistic
FFT
Detect
Inherent
Cycles
Summarization
DWT
DWT
Detect
Turning
Points and
Trends
Analyzing High-frequency
Financial Data: Our Approach
A Case Study
Consider the five minutes compressed tick data for the /$ exchange rate on March 18, 2004
1.0
0.8
Input Data
0.6
1.84
0.4
0.2
0.0
0
25
50
1.83
Feature
1.82
Trend
1.82
0
25
50
75
100
125
150
175
200
225
250
275
Turning
Points
Variance
Change
Cycle
Phrases
Details
1st Phase
2nd Phase
xTrend
3.65e - 6 t + 1.83, 261 < t < 288
2
Downturns
Upturns
5,12,2036,68,and201
Location
164
Period
42
Peaks at
System Output
1.83
Analyzing High-frequency
Financial Data: Our Approach
A Case Study
For prediction,
prediction we use the chief features of the previous day (March 18, 2004), information about the
dominant cycle and trend (summarization), to reproduce the elements of the series for the following
day (March 19, 2004):
Actual
March 19, 2004
1.83
1.83
1.83
Predicted
(seasonal + trend)
March 19, 2004
1.82
1.82
0
25
50
75
100
125
150
175
System Output
200
225
250
Analyzing High-frequency
Financial Data: Our Approach
Conclusions
I.
II.
Multiscale analysis can effectively deconstruct the total series into its constituent
time scales: specific forecasting techniques can be applied to each timescale series
to gain efficiency in forecast
III.
IV.
V.
Questions / Comments