Fast Fourier Transform

Not real sure what you're after here ...

FFT is an algebraic computation method which can be used to analyze a number of selected discrete data points across a selected given time period, outputting the result of the computation process as conventional X/Y graphing coordinates, more or less. It can be applied to just about anything for which event and time can be correlated. As with most math, "garbage in = garbage out". To obtain meaningful information, the selected data points and their relationships to one another would have to be both comprehensive and statistically significant. Just looking at price-vs-time would be an example of a garbage input. For each stock or index considered, numerous data points would be required. The trick would be to figure out just exactly which data points would be statiscally significant; everything from an issue's earning history, market history, relationship to issues within the same and other market segments, etc, to the weather and the general news headlines on a given day could and would be significant data points.

thanx for the info timb

The Kondratieff Wave

also known as the K-Wave

Basically shows from 1790 - present, wholesale prices move in a 52 year cycle.

timberlandko is also right. It's used with regression because when you look at stock chart, the prices move up and down.

The book: Trading Systems and Methods by Perry Kaufman has a good section on cycle analysis using sine and cosine waves (even gets into some basic calculus). as a side note, it also has an astrology section (i.e. tracking the dow using saturn's path and other weird stuff)

Honestly, I think you're better off just looking at what you think the cycle will look like unless you need to write a trading program. If regression could be used to predict stock prices, we'd all be rich.