Dear all: I am searching for a regression model (e.g. y=Xb+e) in which dummy-coded events (to time point t) on different regressors in X exhibit an effect on subsequent responses in the vector y (to time-points t+1, t+2,… t+n). My aim is to estimate how long the memory effect is and how strong a certain event influences subsequent responses (whether the memory decays e.g. linear or exponentially). My first impression is that state-space models are appropriate. However, according to my first understanding in these models regularities in the residuals are explained without using the knowledge of the event history. Or is a simple linear regression model with time-shifted regressors appropriate?
The response vector y includes misses and the events in X are sampled on irregular time points. Many thanks Stefan -- Ist Ihr Browser Vista-kompatibel? Jetzt die neuesten -- ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
