Function that implements the transverse walk sampling strategy to obtain realizations of the posterior distribution of the pumping test parameters.
run_twalk_MCMC(startvalue1, startvalue2, ptest, model, prior.pdf, prior.parameters, iterations = 10000)
| startvalue1 | Numeric vector with the values of the hydraulic parameters used to initialize the chain. |
|---|---|
| startvalue2 | Numeric vector with the values of the hydraulic parameters used to initialize the chain. It must be different than the vector startvalue1. |
| ptest | A pumping test object |
| model | A character string specifying the model that describe the flow during the pumping test |
| prior.pdf | A character vector with the names of the PDF of the hydraulic parameters. Currently only uniform ('unif') and normal ('norm') distributions are supported. |
| prior.parameters | A numeric vector with the parameters of the PDF of the hydraulic parameters. The min and max values of the hydraulic parameters are specified in the case of a uniform distribution. The mean and sigma values of the hydraulic parameters are specified in the case of a normal distribution. |
| iterations | Number of iterations to run the chain. |
This function returns a list with the following entries:
xxp = xxp
logdensity: matrix with the values of the logdensity of the accepted parameters.
acceptance: the acceptance rate
Christen, J. A. & Fox, C. A general purpose sampling algorithm for continuous distributions (the t-walk). Bayesian Analysis, 2010, 5, 2, 263-281.
Other twalk_auxiliary_function functions: blow_kernel,
hop_kernel,
log_ratio_density_blow,
log_ratio_density_hop, psi1,
traverse_kernel, walk_kernel