rtestim (version 1.0.1)

interpolate_rt: Interpolate (or extrapolate) Rt estimates to intermediate design points

Description

Interpolate (or extrapolate) Rt estimates to intermediate design points

Usage

interpolate_rt(object, xout, ...)
# S3 method for cv_poisson_rt
interpolate_rt(object, xout, which_lambda = c("lambda.min", "lambda.1se"), ...)
# S3 method for poisson_rt
interpolate_rt(object, xout, lambda = NULL, ...)

Value

A vector or matrix of interpolated Rt estimates.

Arguments

object

A fitted object produced by estimate_rt() or cv_estimate_rt().

xout

a vector of new positions at which Rt should be produced, but where counts may not have been observed.

...

additional arguments passed to methods.

which_lambda

Select which lambdas from the object to use. If not provided, all Rt's are returned. Note that new lambdas not originally used in the estimation procedure may be provided, but the results will be calculated by linearly interpolating the estimated Rt's.

The strings lambda.min or lambda.1se are allowed to choose either the lambda that minimizes the cross validation score or the largest lambda whose corresponding cross validation score is within 1 standard error of the minimal cross validation score.

lambda

Vector. A user supplied sequence of tuning parameters which determines the balance between data fidelity and smoothness of the estimated Rt; larger lambda results in a smoother estimate. The default, NULL results in an automatic computation based on nlambda, the largest value of lambda that would result in a maximally smooth estimate, and lambda_min_ratio. Supplying a value of lambda overrides this behaviour. It is likely better to supply a decreasing sequence of lambda values than a single (small) value. If supplied, the user-defined lambda sequence is automatically sorted in decreasing order.

Examples

y <- c(1, rpois(100, dnorm(1:100, 50, 15) * 500 + 1))
out <- estimate_rt(y)

# originally estimated at
out$x

# get the Rt at 3 new points (for all estimated lambdas)
int <- interpolate_rt(out, c(10.5, 11.5, 12.5))

# get the Rt at a single value of lambda
interpolate_rt(out, c(10.5, 11.5, 12.5), lambda = out$lambda[20])

y <- c(1, rpois(100, dnorm(1:100, 50, 15) * 500 + 1))
out <- estimate_rt(y, nsol = 10)
interpolate_rt(out, xout = c(1.5, 2.5))