|
PureBytes Links
Trading Reference Links
|
Thanks for the emails
Volatility clustering is another term for conditional hetroskedasticity.
This means that "aberrant outliers", (which traditionally trained stats
people want to get rid of because it means "bad data"), appear in clusters.
This "bubble" characteristic of price data or time series data is most
interesting. <G>
Some have asked about Constance Brown's book and Kase's book ... you might
want to check out Blake LeBaron's article "Volatility Persistence and
Apparent Scaling Laws in Finance" July '99. It reviews the history of
scaling (i.e., using different time scales) etc. which is popular in more
advanced technical analysis. See the following for the abstract and download
site
http://stanley.feldberg.brandeis.edu/~blebaron/id23.htm
He asks some very interesting questions and casts doubt on some popular TA
ideas.
There are ideas here that will take you into the area of econophysics and
especially the Olsen and Associates group at www.olsen.ch This group has
done a lot for high speed time series data research (i.e., 1 and 5 min
data).
You didn't tell me that the Demetra output file for a time series results
and diagnostics was 40+ pages. What a monster program. Thanks to all the
good European folk who paid their taxes to provide this free program and
service.
Best regards
Walter
|