Engineers Please advise

I am doing a DIY of this type of sinewave anaylsis

www.cycles.cc
www.techsignal.com

This trader is very successful with this stuff: http://www.charlesnenner.com/index.php

 

Digs,

Autoregressive Processes and binomial filter smoothing is what you are looking to do.

cosine will get you there.

best

 

Wave,

Thanks, how does one use Cosine with three data sets of sinewave ???

 

Digs,

I had a spreadsheet at one point in time that did exactly what you are trying to do. From memory, I remember the steps as being something along these lines:

1. Take the running cosine of all three datasets values

2. Multiply #1 by (PI()/rate)+PI()*1.1)

3. Multiply #2 by amplitude + median of all values in #1

hope that helps.

 

Good Luck with your approach.

Have a look at the FFT pitfalls here for a more 'engineering' oriented viewpoint.
http://intelligenttradingtech.blogspot.com/2010/02/fft-of-time-series-promises-and.html

 

Wave,

That may be more maths than I need.

So what was your conclusion of sinewave studies ..

cheers

 

I also have this, and its not bad..

http://emini-watch.com/products/hilbert-sine-wave

 

Not the holy grail. After much research, I just use price.