http://ai-quant.com/2008/02/17/fourier-transforms-and-financial-time-series-data/ Signal Processing applied to the financial markets. Sat, 17 May 2008 00:44:31 +0000 http://wordpress.org/?v=2.3.1 http://ai-quant.com/2008/02/17/fourier-transforms-and-financial-time-series-data/#comment-527 aiquant Mon, 10 Mar 2008 08:00:27 +0000 http://ai-quant.com/2008/02/17/fourier-transforms-and-financial-time-series-data/#comment-527 Yes, the cycle needs to be fixed and time invariant over the observation window - which is usually not the case. The Hilbert transform provides a mechanism of determining the cycle length using as little as four bars of price data and is well suited to an automated trading environment. Yes, the cycle needs to be fixed and time invariant over the observation window - which is usually not the case. The Hilbert transform provides a mechanism of determining the cycle length using as little as four bars of price data and is well suited to an automated trading environment.

]]> http://ai-quant.com/2008/02/17/fourier-transforms-and-financial-time-series-data/#comment-509 Juan Carlos Christensen Sun, 09 Mar 2008 00:35:56 +0000 http://ai-quant.com/2008/02/17/fourier-transforms-and-financial-time-series-data/#comment-509 FT needs at least one complete cycle, isn't that true? I mean, if you have a clean sine signal, going from phase 0 to phase 180 (half cycle) the FT would not measure the cycle correctly (while some other tools like MEM would). Please correct me if im wrong. FT needs at least one complete cycle, isn’t that true? I mean, if you have a clean sine signal, going from phase 0 to phase 180 (half cycle) the FT would not measure the cycle correctly (while some other tools like MEM would). Please correct me if im wrong.

]]>