An Application of the Continuous Wavelet Transform to Financial Time Series
(2018) EITM01 20082Department of Electrical and Information Technology
- Abstract
- Wavelet theory, which shares fundamental concepts with windowed Fourier analysis, introduces the notion of scale in an effort to aid in joint time-frequency analysis. Having century-old roots, much of the essential research on the subject of wavelets was conducted during the 1970s and 1980s. Despite being a rather young toolset, wavelets have shown to be very useful when studying signals with transient, non-stationary, characteristics.
This thesis focuses on the continuous wavelet transform (CWT) in the one-dimensional case from a practical implementation standpoint. It also contains sections on wavelet history, development, and the theoretical fundamentals.
The presented implementation contains a computer software with a graphical... (More) - Wavelet theory, which shares fundamental concepts with windowed Fourier analysis, introduces the notion of scale in an effort to aid in joint time-frequency analysis. Having century-old roots, much of the essential research on the subject of wavelets was conducted during the 1970s and 1980s. Despite being a rather young toolset, wavelets have shown to be very useful when studying signals with transient, non-stationary, characteristics.
This thesis focuses on the continuous wavelet transform (CWT) in the one-dimensional case from a practical implementation standpoint. It also contains sections on wavelet history, development, and the theoretical fundamentals.
The presented implementation contains a computer software with a graphical user interface that was developed in the context of financial trading in the currency markets. More specifically, the implementation contains a C++ based code library developed to expose an application programming interface (API) that is called from a retail desktop forex trading software where it can aid in market analysis visualization. (Less) - Popular Abstract
- A Look at Volatility in Financial Time Series Through Wavelet Analysis
Each year, billions of dollars are spent finding an edge in the financial markets. At the forefront of this pursuit is Wall Street's elite -- the Quants -- bringing physics, signal processing and engineering to the world of finance and economics. With this backdrop of high-stake applied research and evaluation of new tools and theories, we develop a trading software indicator using wavelet analysis and bring it into practice by exploring volatility phenomenon in the currency markets.
In the field of signal processing, it is well known that the legendary Fourier transform help us move between the often empirically observed time domain and the theoretically... (More) - A Look at Volatility in Financial Time Series Through Wavelet Analysis
Each year, billions of dollars are spent finding an edge in the financial markets. At the forefront of this pursuit is Wall Street's elite -- the Quants -- bringing physics, signal processing and engineering to the world of finance and economics. With this backdrop of high-stake applied research and evaluation of new tools and theories, we develop a trading software indicator using wavelet analysis and bring it into practice by exploring volatility phenomenon in the currency markets.
In the field of signal processing, it is well known that the legendary Fourier transform help us move between the often empirically observed time domain and the theoretically fascinating frequency domain. By studying the transformed signal, we can easily reveal the set of frequencies it is composed of. However, operating globally on the signal content, the Fourier transform does not lend itself to exploration of when certain frequency events occur in time. In other words, traditional Fourier analysis falls short when used in a non-stationary, transient settings and this is where wavelets come into play. By introducing the notion of scale as a proxy to frequency, the wavelet framework approaches the problem of telling when an exact frequency occurs in time; a dilemma with its roots in the Heisenberg uncertainty principle.
One setting where these types of transient signals -- or time series -- are prevalent, is the world of finance and economics. In this thesis work, we implemented the highly theoretical wavelet framework in a desktop trading application for the currency markets and used it to explore market movements (or volatility in a broad sense). Volatility analysis is an invaluable concept for many traders, institutional and retail alike. Front-running in the sense of bringing wavelet based volatility analysis straight into the retail currency trading world, the implementation resulted in a C++ library, which was consumed and visualized in the form of a market indicator. The indicator was based on the scalogram output of the continuous wavelet transform. Similar to a heatmap, it can easily identify parts of the time series containing abrupt changes or clusters of high volatility. While implementing the technical solution we stumbled upon several challenges in bridging theory and practice. Despite this, both the general software library and practical indicator is fully functional and ready to be shared with an enthusiastic trading community for further exploration and battle testing. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/8938181
- author
- Eliasson, Klas LU
- supervisor
-
- Stefan Höst LU
- organization
- course
- EITM01 20082
- year
- 2018
- type
- H2 - Master's Degree (Two Years)
- subject
- keywords
- Wavelets, Continuous Wavelet Transform, CWT, Financial Time Series, Currency Trading
- report number
- LU/LTH-EIT 2018-625
- language
- English
- id
- 8938181
- date added to LUP
- 2018-04-13 12:00:46
- date last changed
- 2018-04-13 12:00:46
@misc{8938181,
abstract = {{Wavelet theory, which shares fundamental concepts with windowed Fourier analysis, introduces the notion of scale in an effort to aid in joint time-frequency analysis. Having century-old roots, much of the essential research on the subject of wavelets was conducted during the 1970s and 1980s. Despite being a rather young toolset, wavelets have shown to be very useful when studying signals with transient, non-stationary, characteristics.
This thesis focuses on the continuous wavelet transform (CWT) in the one-dimensional case from a practical implementation standpoint. It also contains sections on wavelet history, development, and the theoretical fundamentals.
The presented implementation contains a computer software with a graphical user interface that was developed in the context of financial trading in the currency markets. More specifically, the implementation contains a C++ based code library developed to expose an application programming interface (API) that is called from a retail desktop forex trading software where it can aid in market analysis visualization.}},
author = {{Eliasson, Klas}},
language = {{eng}},
note = {{Student Paper}},
title = {{An Application of the Continuous Wavelet Transform to Financial Time Series}},
year = {{2018}},
}