genmax: New version that sends down arrays instead of pointers!!

Description Usage

Description

# 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.

Usage

 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
)

Auburngrads/SMRD documentation built on Sept. 14, 2020, 2:21 a.m.