[R-sig-finance] recursive penalized regression
Vadim Ogranovich
vograno at evafunds.com
Sat Jan 15 00:30:11 CET 2005
Hi,
I have a question which I believe has been discussed a number of times
on both r-sig-finance and r-help, but somehow I am not able to locate
relevant messages. My apologies for bringing it up again.
I want to do a univariate no frills autoregression. The major
non-standard requirements are:
1. When estimating model parameters at time t the algorithm can only use
data up to time t.
2. The weights of the past observations should decay with time, e.g
exponentially
3. ability to apply some penalty, e.g. L2 (ridge), L1 (lasso), etc., to
model coefficients.
I think that all of these fall under the Kalman Filter title (except
probably for lasso), but a) search for "Kalman Filter" turns up zillions
of messages and it's hard for me to tell their relevance or suitability
for the task at hand and b) maybe some relevant code hides under some
other title. For example I remember reading about a function which
satisfies 1., but as far as I remember it had nothing to do with the
Kalman filters .
Thanks in advance for your help,
Vadim
[[alternative HTML version deleted]]
More information about the R-sig-finance
mailing list