Penalized Least Squares for Smoothing Financial Time Series

Adrian Letchford, Junbin Gao, Lihong Zheng

Research output: Book chapter/Published conference paperConference paper

57 Downloads (Pure)


Modeling of financial time series data by methods of artificialintelligence is difficult because of the extremely noisy nature of thedata. A common and simple form of filter to reduce the noise originatedin signal processing, the finite impulse response (FIR) filter. There areseveral of these noise reduction methods used throughout the financialinstrument trading community. The major issue with these filters is thedelay between the filtered data and the noisy data. This delay only increasesas more noise reduction is desired. In the present marketplace,where investors are competing for quality and timely information, thisdelay can be a hindrance. This paper proposes a new FIR filter derivedwith the aim of maximizing the level of noise reduction and minimizingthe delay. The model is modified from the old problem of time seriesgraduation by penalized least squares. Comparison between five differentmethods has been done and experiment results have shown that ourmethod is significantly superior to the alternatives in both delay andsmoothness over short and middle range delay periods.
Original languageEnglish
Title of host publicationAI 2011
Subtitle of host publicationAdvances in Artificial Intelligence
Place of PublicationGermany
Number of pages10
Publication statusPublished - 2011
EventAustralian Joint Conference on Artificial Intelligence - Perth, Australia
Duration: 05 Dec 201108 Dec 2011

Publication series

ISSN (Print)0302-9743


ConferenceAustralian Joint Conference on Artificial Intelligence

Fingerprint Dive into the research topics of 'Penalized Least Squares for Smoothing Financial Time Series'. Together they form a unique fingerprint.

  • Cite this

    Letchford, A., Gao, J., & Zheng, L. (2011). Penalized Least Squares for Smoothing Financial Time Series. In AI 2011: Advances in Artificial Intelligence (Vol. 7106, pp. 72-81). Springer.