toys               package:calibrator               R Documentation

_T_o_y _d_a_t_a_s_e_t_s

_D_e_s_c_r_i_p_t_i_o_n:

     Toy datasets that illustrate the package.

_U_s_a_g_e:

       data(toys)
       D1.toy
       D2.toy
       d.toy
       phi.toy
       theta.toy
       V.toy
       X.dist.toy

_F_o_r_m_a_t:

     The 'D1.toy' matrix is 8 rows of code run points, with five
     columns.  The first two columns are the lat and long and the next
     three are parameter values.

     The 'D2.toy' matrix is five rows of observations on two variables,
     'x' and 'y' which are styled ``latitude and longitude''.

     'd.toy' is the "data" vector consisting of length 13: elements 1-8
     are code runs and elements 9-13 are observations.

     'theta.toy' is a vector of length three that is a working example
     of theta.  The parameters are designed to work with
     'computer.model()'.

     't.vec.toy' is a matrix of eight rows and three columns.  Each row
     specifies a value for theta.  The eight rows correspond to  eight
     code runs.

     'x.toy' and 'x.toy2' are vectors of length two that gives a sample
     point at which observations may be made (or the code run). The
     gloss of the two elements is latitude and longitude.

     'x.vec' is a matrix whose rows are reasonable x values but _not_
     those in 'D2.toy'.

     'y.toy' is a vector of length eight.  Each element corresponds to
     the output from a code run at each of the rows of 'D1.toy'.

     'z.toy' is a vector of length five.  Each element corresponds to a
     measurement at each of the rows of 'D2.toy'.

     'V.toy' is a five by five variance-covariance matrix for the toy
     datasets.

     'X.dist.toy' is a toy example of a distribution of 'X' for use in
     calibrated uncertainty analysis, section 4.2.

     *Brief description of toy functions fully documented under their
     own manpage*

     Function 'create.new.toy.datasets()' creates new toy datasets with
     any number of observations and code runs.

     Function 'E.theta.toy()' returns expectation of 'H(D)' with
     respect to theta; 'Edash.theta.toy()' returns expectation with
     respect to Edash.

     Function 'extractor.toy()'  extracts 'x.star.toy' and 't.vec.toy'
     from 'D2'; toy example needed because the extraction differs from
     case to case.

     Function 'H1.toy()' applies basis functions to rows of 'D1' and
     'D2'

     Function 'phi.fun.toy()' creates a hyperparameter object such as
     'phi.toy' in a form suitable for passing to the other functions in
     the library.

     Function 'phi.change.toy()' modifies the hyperparameter object.

     *See the helpfiles listed in the "see also" section below*

_D_e_t_a_i_l_s:

     All toy datasets are documented here.  There are also several toy
     functions that are needed for a toy problem; these are documented
     separately (they are too diverse to document fully in a single
     manpage).  Nevertheless a terse summary  for each toy function is
     provided on this page.  All toy functions in the package are
     listed under "See Also".

_A_u_t_h_o_r(_s):

     Robin K. S. Hankin

_R_e_f_e_r_e_n_c_e_s:

     M. C. Kennedy and A. O'Hagan 2001. "Bayesian calibration of
     computer models".  Journal of the Royal Statistical Society B,
     63(3) pp425-464

     M. C. Kennedy and A. O'Hagan 2001.  "Supplementary details on
     Bayesian calibration of computer models", Internal report,
     University of Sheffield.  Available at <URL:
     http://www.shef.ac.uk/~st1ao/ps/calsup.ps>

     R. K. S. Hankin 2005. "Introducing BACCO, an R bundle for Bayesian
     analysis of computer code output", Journal of Statistical
     Software, 14(16)

_S_e_e _A_l_s_o:

     'create.new.toy.datasets', 'E.theta.toy', 'extractor.toy',
     'H1.toy', 'phi.fun.toy', 'stage1'

_E_x_a_m_p_l_e_s:

     data(toys)
     D1.toy
     extractor.toy(D1.toy)

     D2.fun(theta=theta.toy , D2=D2.toy)
     D2.fun(theta=theta.toy,D2=D2.toy[1,,drop=FALSE])

     library("emulator")
     corr.matrix(D1.toy,scales=rep(1,5))
     corr.matrix(D1.toy, pos.def.matrix=diag(5))

