CATS Seminars restarting soon

For discussion of how to use the Cycles Research Institute' own Cycles Analysis & Time Series software.

CATS Seminars restarting soon

Postby RayTomes » 21 Sep 2010, 01:20

I am in the process of getting a seminar tool set up so that we can hold CATS seminars with many people present at once. Please keep an eye on this thread as an announcement will be made shortly. I expect to hold the first seminar about a week from today.

The topic will be taking a time series and determining the significant cycles present in it and reconstructing estimates of the series from the cycles so as to also produce projections. Always remember that it has been noted that forecasting is a difficult thing to do, especially about the future. :-)
User avatar
RayTomes
 
Posts: 128
Joined: 02 Aug 2010, 23:24
Location: Auckland, New Zealand

Re: CATS Seminars restarting soon

Postby RayTomes » 20 May 2011, 12:57

stuhiggins wrote:...
I have watched the four vimeo presentations on the software yesterday, and would like to ask the following questions :
Copper paper
how do i remove the linear trend of prices before calculating the spectrum (per your copper paper)(detrend)
how can i add several cycle frequencies together to get a blended cycle (per your copper paper)
what do you mean by a smoother 3 cycle filter of 155m period (per copper chart in paper) (how can I create the chart)
what do you mean by a smoother 25 month mov avg less 77 month mov avg on 77.65m period (per copper chart in paper) (how can I create the chart)
General use
how can I extract data from the system (for example, log prices, specturm, line of best fir)
how many series can i perform pca on
how can you perform seasonality analysis, and how can this structure be incorporated into the line of best fit
if i have shortened the dates of the data for a chart, how can i lengthen them again
a video on the use of lambda and also principal factor analysis would be great
Data
I would like to locate data on planetary alignments (saturn v uranus v jupiter v sun) - do you know of any good sources.
is there a comprehensive list of the data contents in cats folders v respective numbers
It looks like you are interested in planetary forces. Not sure if you have seen this: http://www.esc-web.org/papers/potsdam_2 ... nathan.pdf
Appreciate your efforts in bringing the software to the general public.


Sorry that I have not been so active lately. Hopefully CATS training program will get started again before too long.

Linear trend can be removed by regression against a variable made with I function. There is a macro Udetr for detrending.

To add several cycles together, use CATS commands "nn.nn Usc" for each period nn.nn to create sine and cosine components for each cycle, the put the detrended series on the stack and use Rn where n is 2 x no. of periods. This will make a regression equation. Use Mn to calculate estimates.

You create smoothed mov avg chart by , e.g. Fn @25 Fn @77 - to get the difference between moving averages. Usually I use centred mov avgs so Fn @25 :-13 Fn @77 :-39 -

What is "pca"? Principle Components Analysis? You can use 100 series for factor analysis in present CATS version. Next version may be higher I think.

Seasonal trend can be done with W12 (for monthly data) to get average shape (best to detrend first).

You can lengthen dates with Dstart,end to set start and end dates.

Have fun!
Ray
User avatar
RayTomes
 
Posts: 128
Joined: 02 Aug 2010, 23:24
Location: Auckland, New Zealand


Return to CATS Software

Who is online

Users browsing this forum: No registered users and 2 guests

cron