Lags vectors and covariates correctly so that an autoregressive model
can be estimated by regression.
Usage
make.lags(x, lags, cov,nobs=3500)
Arguments
x
|
Vector or matrix representing a univariate or multivariate time series.
(rows are assumed to index time)
|
lags
|
Vector of time delays used in reconstruction.
|
nobs
|
Maximum length of time series.
|
cov
|
A vector or matrix of covariates that will be matched with the times for
the independent variable
|
Description
This function is used to create the appropriate data structure for
a nonlinear autoregressive process of the form X_t = F(X_t-1) + e_t.Value
x
|
Matrix of lagged values of the time series, independent variables.
The covariates are the last columns of this matrix
|
y
|
Vector of time series values, dependent variables.
|
nvar
|
Number of variables or dimension of x matrix.
|
lags
|
Time delays used in constructing the x matrix.
|
start
|
Observation number of univariate time series used for the start of the
y vector.
|
end
|
Observation number of univariate time series used for the end of the
y vector.
|
skip
|
Information about which columns of the returned X matrix are covariates.
|
See Also
nnreg, rosslerExamples
make.lags(rossler.state[,1],c(1,2,3)) -> data
# create
# 3-d time delay vector model of the x variable of rossler system.
nnreg(data$x,data$y,5,5) -> fit # fit time series model using nnreg.
# fitting a state space model to the rossler state vector
# only one lag is needed in this case.
make.lags(rossler.state, lags=c(1))-> data
nnreg( data$x, data$y[,1], 5,5)-> fit1
nnreg( data$x, data$y[,2], 5,5)-> fit2
nnreg( data$x, data$y[,3], 5,5)-> fit3