# kmod (I) model code (usual list 1-5) # ————————————————————- # 0 distribution kdist with one or more parameters related to # explanatory variables # 1 same as 0 with an lfp parameter # 2 same as 0 with a doa parameter # 3 steady-state model with distribution kdist plus exponential # competing risk # 4 proportional hazards model for distribution kdist plus # explanatory variables for the power (or other) parameters # # kdist (I) distribution number (usual list 1-12) # >100 is a user specified distribution # # theta(nparm) (O) double precision vector returning ml estimates # of model parameters and values of any # fixed parameters # # thetas(nparm) (O) double precision vector returning ml # scaled estimates # of model parameters and (scaled) values of # any fixed parameters # data scaled according to centered and scaled x # # kodet(nparm) (O) integer vector of parameter codes indicating range # 0 fixed # 1 unrestricted # 2 positive # 3 0-1 # # # ifix(nparm) (I) an integer vector of 0's and 1's to indicate which # parameters are to be fixed (1 for fixed) # # nparm (O) number of model parameters # nparm is the sum of the parameters in all of the # regression relationships plus any model parameters # that do not depend on explanatory variables # If this number is needed above before calling # genmax (to allocate parameter space) it can be # gotten from a call to gensiz # the model stuff (kmod,kdist,kparv,nrvar,mrelat,nrelat) # has been set. # # npard (O) number of model parameters # npard is the number of parameters in the distributional # model (i.e., for a particular combination of the # explanatory variables # If this number is needed above before calling # genmax (to allocate parameter space) it can be # gotten from a call to gensiz # the model stuff (kmod,kdist,kparv,nrvar,mrelat,nrelat) # has been set. # # y (I) real matrix of response times . # transformed times may be returned # # ncoly (I) number of columns in the observation matrix (1 or 2) # # nrownw (I) number of rows in the data matrix # # x (I) real matrix x of explanatory variables # which must have a # column of ones in the first column (even if it # is not used). # # ***** note that upon return, x may not be in its original # form due to scaling or other standardization. # # ncolx (I) number of columns in x, not including the required col # of ones (because we think of the ones as column zero) # # codes (I) real vector of censor codes # # weight (I) real vector of observation weights # # ty (I) real matrix of truncation times # # ncolty (I) number of columns in the matrix of truncation times # (0, 1, or 2) # # tc (I) real vector of truncation codes # # kprint (I) print code 0-none 1-minimal 2-usual 3-light debug # 4 and higher provides more debug output # # # kparv (I) integer vector giving parameter # numbers having regression # relationships # # nrvar (I) integer vector. nrvar(i) gives the number of columns # of x that are in the relationship for the # parameter specified in kparv(i) # # mrelat(mnrvar,nrelat) (I) integer matrix # in which entry i,j gives the col of x # for the ith term of the jth relationship. i=1,nrvar(j), # j=1,nrelat # # nrelat (I) number of regression relationships # # mnrvar (I) integer vector. nrvar(i) gives the number of columns # # xlogl (O) double precision scaler returning value of the # log likelihood # # yhat (O) real vector (matrix) of fitted values (same size as y) # # resid (O) real vector (matrix) of residuals (same size as y) # # vcv(nparm,nparm) (O) double precision matrix returning estimated matrix of # variances and covariances of the ml estimates # # vcvs(nparm,nparm)(O) double precision matrix returning estimated # matrix of variances and covariances of the # ml scaled estimates # # r(nparm,nparm) (O) double precision matrix returning estimated # correlation matrix from the vcv matrix # # start (I–opt) optional vector of starting parameter values # (ignored if lstar=0) # # lstar (I) =1 is start values are in start # =0 if automatic start values are # to be computed velow # # ilabp (O) integer vector of parameter labels (length 8*nparm) # in this new version we copy over to stack and pass # pointers below # # ilabd (O) integer vector of distribution parmameters # (length 8*nparm) # in this new version we copy over to stack and pass # pointers below # # ier code meaning # ——————————————————————- # third digit: 0 no optimization errors detected # 1 likelihood shape caused problems with the powell alg # 2 convergence criterion not met after maximum number # or iterations # # second digit: 0 first derivatives of likelihood small # 1 first derivatives of the loglikelihood too large # # first digit: 0 estimated fisher info matrix inverted successfully # 1 estimated fisher info matrix appears to be singular # #note: all pointers sent down in the argument list are pointing to the # rs(.) stack in labeled common. # # fortran typing conventions are generally followed for typing # except that double precision is used except when data is to # be taken from the worksheet.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 | GENMAX(
kmod,
kdist,
theta,
thetas,
kodet,
ifix,
nparm,
npard,
y,
ncoly,
nrownw,
x,
ncolx,
codes,
weight,
ty,
ncolty,
tcodes,
kprint,
kparv,
nrvar,
mrelat,
nrelat,
mnrvar,
xlogl,
yhat,
resid,
vcvs,
vcv,
r,
start,
lstar,
conlev,
ilabp,
ilabd,
ier,
nxd,
intd,
ipxcd,
irelad,
fstder,
nregr,
kcentr,
kpoint,
ifit,
kgtall,
llog,
kmodp,
maxit,
pest,
epsx,
npardm,
nnum,
kparm,
iup,
nterd,
maxpd,
pfail,
kmccde,
nstart,
maxmsd,
tol,
lsd,
pchmax
)
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.