| %tt% {STAR} | R Documentation |
Performs time transformation using a gssanova fit. If the model
is correct, the result of the transformation should be a Poisson
process with rate 1.
gssObj %tt% dataFrame
gssObj |
a gssanova or a gssanova0 object. |
dataFrame |
a data.frame with variables corresponding to
the ones used in the gssanova call giving rise to gssObj. |
The binary operator applies
predict.ssanova with the left side as the first
argument and the right side as the second argument. The right side
(dataFrame) must therefore contain the variables included in
the formula used in the call giving rise to gssObj. The
result of the predict method call is then transformed with an
inverse logistic function or with an exponential (depending on the
family argument, "binomial" or "poisson", used in
the previous gssanova call). The cumulative sum is
computed, that is, the integrated conditional intensity, and its value
at the events times is returned as a CountingProcessSamplePath object.
A CountingProcessSamplePath object.
Christophe Pouzat christophe.pouzat@gmail.com
Gu C. (2002) Smoothing Spline ANOVA Models. Springer.
Brillinger, D. R. (1988) Maximum likelihood analysis of spike trains of interacting nerve cells. Biol. Cybern. 59: 189–200.
Brown, E. N., Barbieri, R., Ventura, V., Kass, R. E. and Frank, L. M. (2002) The time-rescaling theorem and its application to neural spike train data analysis. Neural Computation 14: 325-346.
Ogata, Yosihiko (1988) Statistical Models for Earthquake Occurrences and Residual Analysis for Point Processes. Journal of the American Statistical Association 83: 9-27.
gssanova,
predict.ssanova,
mkGLMdf,
mkCPSP,
summary.CountingProcessSamplePath
## Not run: ## load e060517spont data set data(e060517spont) ## make a data frame using a 2 ms bin width e060517spontDF <- mkGLMdf(e060517spont,0.002,0,60) ## Keep data relevant to neuron 3 e060517spontDFn3 <- e060517spontDF[e060517spontDF$neuron == "3",] ## Split data in an "early" and a "late" part e060517spontDFn3e <- e060517spontDFn3[e060517spontDFn3$time <= 30,] e060517spontDFn3l <- e060517spontDFn3[e060517spontDFn3$time > 30,] ## fit the late part with a nonparametric renewal model e060517spontDFn3lGF <- gssanova(event ~ lN.3, data=e060517spontDFn3l,family="binomial") ## transform the time of the early part e060517spont.n3e.tt <- e060517spontDFn3lGF %tt% e060517spontDFn3e ## Test the goodness of fit e060517spont.n3e.tt summary(e060517spont.n3e.tt) plot(summary(e060517spont.n3e.tt)) ## End(Not run)