DtmVic – Parameters of the Command File

(For advanced users)

For most usual applications, these parameters are automatically generated from the button “Create a command file" of the main menu. However, the command file could also be edited in a subsequent phase (change of parameters values, replacement of some default values, addition of a new step, etc.)

List of Steps
  afcor (correspondence analysis [CA])   aplum (CA of lexical tables)   ardat (building archive dictionary and data)   artex (building archive textual data)   aspar (Direct CA of texts)   clair (description of axes in textual analysis)   corda (concordances)   corte (correcting texts)   decat (description of categories of a nominal variable)   decla (description of clusters - after steps 'semis' or 'recip')   defac (description of principal axes)   ipfit (iterative proportional fitting)   mocar (characteristic words)   motex (table categories x texts)   multm (multiple correspondence analysis)   numer (numeric coding of texts)   parti (cuts a dendrogram and optimizes the partition)   permu (permutation of variables (columns)and individuals (rows) )   polex (positionning words on principal space)   posit (positionning categories in textual principal space)   prico (principal components analysis)   recar (characteristic responses)   recip (hierarchical clustering: reciprocal neighbours)   regre (regression and analysis of variance)   segme (computing repeated segments)   selec (selecting active and illustrative variables and individuals)   selox (selecting an open question)   semis (partitioning through k-means method)   setex (setting a new threshold to select words)   splum (selecting text categories in CA of lexical table)   stats (basic statistics for numerical and nominal variables)   table (cross tabulations with filters)   talex (stacked lexical tables)




afcor (correspondence analysis [CA])
  [Return to List of Steps]
Parameters 
  naxe:   number of requested axes.
  nvedi:  number of eigenvalues (histogram).
  ledin:  printing the coordinates of the 
            individuals (0=no, 1=act, 2=tot).
  naxed:  number of printed coordinates.
  niter:  number of iterations (case of iterative 
           algorithm) (0=no).
  nictl:  number of "controlled iterations" (0=no).
  nbase:  dimension of the approximation subspace.
  nsimu:  number of bootstrap replication 
           (0 = no bootstrap, max = 30)
  nboot:  type of bootstrap replication 
          ( 0 = "total bootstrap", 1= partial
            bootstrap [default])
  ndata:  printing the replicated data tables 
          (1 = yes, 0 = no [default])

  

aplum (CA of lexical tables)
  [Return to List of Steps]
Parameters 
  naxe:    number of principal coordinates.
  ledit:   parameter for printing the rows 
            (0=no, 1=yes).
  ngraf:   number of requested graphics
  npage:   number of pages of the graphics
  nrows:   number of lines of the graphics
  nsimu:   number of bootstrap replication 
            (0 = no bootstrap, max = 30)
  nboot:   type of bootstrap replication 
             0 = "total bootstrap", 
                 [bootstrapping the words]
             1 = partial bootstrap [default]
                 [bootstrapping the words]
             2 = does not apply here
             3 = total specific bootstrap
                 [bootstrapping the respondents]
             4 = partial specific bootstrap
                 [bootstrapping the respondents]
   ndata:   printing the replicated data tables 
            (1 = yes, 0 = no [default])


ardat (building archive dictionary and data)
  [Return to List of Steps]
Parameters 
  nqexa:   number of variables in the  dictionary.
  nxmod:   maximal number of categories 
            for a variable.
  ledit:   printing mode of the dictionary 
            (0=no, 1=short, 2=long)
  nidi:    number of groups of 4 characters 
            (identifier of individuals) (0=no).
  niexa:   number of individuals in file ndonz.
  test:    value given to missing numerical data (default=.999).
  nedit:   number of individuals in a check 
            printout (0=no).
 
artex (building archive textual data)
  [Return to List of Steps]
Parameters 
  ityp:    type of textual data file ntexz 
             (tex=1, enq =2)
  nbqt:    number of questions per individual
  nlig:    number of responses to be printed for 
             each text (0 = only the titles)
  lirep:   indicates implicit end of a response 
             (1= 1 line (row)/1 response)


aspar ( Direct CA of texts)
  [Return to List of Steps]
Parameters 
  naxe:    number of requested principal coordinates
  ledit:   printing the responses 
             (0 = no, 1 = coor, 2 = tot)
  ngraf:   number of requested graphics
  npage:   number of pages of the graphics
  nrows:   number of lines of the graphics
  nbase:   dimension of the approximation space 
             (0= main core diagonalization)
  niter:   number of iterations (if nbase > 0)


clair (description of axes in textual analysis)
  [Return to List of Steps]
Parameters 
  naxe:    number of described axes
  lign:    simultaneous listing of the rows 
             (0=no, 1=yes).
  nmax:    number of printed "extreme elements" 
             for each principal axis
 

corda (concordances)
  [Return to List of Steps]
Parameters 
  ledit:   printing identifiers of individuals
             (0 = no printing, 1 = identifiers of
              respondents are printed, default = 0) 
  lspb:    use of a file nspb  (0=no, 1=yes)
        --- key-word of headings : forme end


corte (correcting texts)
  [Return to List of Steps]
Parameters 
  ledit:   printing of words  (0=no, 1=nspc, 2=tot).
  lclas:   printing sorted words (0=no, 1=yes).
          --- key-word of headings : delet equiv end


decat (description of categories of a nominal variable)
  [Return to List of Steps]
Parameters     
  ednom:   printing the tables (classes * questions) 
             (0=no).
  lpnom:   describing partition with questions 
             (0=no, 1=yes).
  cnomi:   describing classes with questions (0=no).
  pmoda:   describing partition with categories (0=no).
  cmoda:   describing classes with categories (0=no).
  pcmin:   minimum relative weight (%) for a category.
  lsupr:   characteristic category if
             %(cat./class) > %(cat./total) 
             (0=no, 1=yes).
  pfreq:   describing partition with frequencies (0=no).
  lpfre:   printing thetable (classes * frequencies) 
             (0=no, 1=act, 2=tot).
  cfreq:   describing classes with frequencies (0=no).
  edcon:   describing partition with numerical variables  
             (0=no, 1=yes).
  ccont:   describing classes with numerical variables 
             (0=no).


decla (description of clusters - after steps 'semis' or 'recip')
  [Return to List of Steps]
Parameters     
  ednom:   printing the tables (classes * questions) 
             (0=no).
  lpnom:   describing partition with questions 
             (0=no, 1=yes).
  cnomi:   describing classes with questions (0=no).
  pmoda:   describing partition with categories (0=no).
  cmoda:   describing classes with categories (0=no).
  pcmin:   minimum relative ( % ) weight for a category.
  lsupr:   characteristic category if 
             %(cat./class) > %(cat./total)(0=no,1=yes).
  pfreq:   describing partition with frequencies (0=no).
  lpfre:   printing  the table (classes * frequencies) 
            (0=no, 1=act, 2=tot).
  cfreq:   describing classes with frequencies (0=no).
  edcon:   describing partition with numerical variables  
            (0=no, 1=yes).
  ccont:   describing classes with numerical variables 
             (0=no).
  paxe:    describing partition with axes (0=no).
  caxe:    describing classes with axes(0=no).
  ledin:   printing correspondence classes*individuals 
           (0=no,1=compo, 2=affec, 3=tot).


defac (description of principal axes)
  [Return to List of Steps]
Parameters 
  seuil:   threshold for selecting most significant items.
  lcrim:   criterion of selection for the categories 
             ( coordinate or test-value)
  vtmin:   test-value threshold for selecting categories.


ipfit (Iterative Proportional FITting)
  [Return to List of Steps]
Parameters 
  niter:     Maximum number of iterations                                                   
  acdif:     Maximum difference between required and obtained percentages                   
  liste:      Printing the individual weights (0=no, 1=yes).                                 
  ledit:      Printing the future provisional dictionnary (0=no , 1=short, 2=long).                      
  nxvrd:      Maximal number of categorical variables used for the iterative fitting         
  key word: end (after lists of target [or: required] percentages)

mocar (characteristic words and responses)
  [Return to List of Steps]
Parameters 
  nomot:   number of charac. words  for each text.
  norep:  number of charac. responses for each text.


motex (table categories x texts)
  [Return to List of Steps]
Parameters 
  nvsel:   index of the categorical variable defining
             the groupings of texts
  ledit:   parameter for printing the table words*texts
             (0=no, 1=yes).


multm (multiple correspondence analysis [MCA])
  [Return to List of Steps]
Parameters 
  naxe:       number of requested axes.
  pcmin:   threshold for "cleaning" of the 
             active categories (in percent).
  lburt:   printing the Burt table 
             (0=no, 1=mass, 2=tot, 3=prof).
  nvedi:   number of eigenvalues (histogram).
  naxed:   number of printed coordinates.
  ledin:   printing the coordinates of individuals 
            (0=no, 1=act, 2=tot).
  ledco:   printing the correlations variable-axers 
             (0=no, 1=yes).
  nsimu:  number of bootstrap replication 
            (1 = no bootstrap)
  nboot:    (1 = partial bootstrap [default])
            (0 = total bootstrap. In such a case,
             the "cleaning of rare categories"
            [parameter pcmin above] is not allowed)


numer (numeric coding of texts)
  [Return to List of Steps]
Parameters 
  coef:    technical coefficient 
            (frequencies of word-lengths)
            (default value = 1; put coef=10 for large corpora).
  nseu:    frequency threshold of the kept words
  nxlet:   maximum number  of letters per word.
  nxmax:   maximum number  of distinct words.
  nxsig:   maximum number of first letters of word.
  ledit:   printing the words (0=no, 1=alpha, 2=tri, 3=tot).
  nmomi:   threshold for the length of words
      --- key-word of headings
           weak strong accent histo end


parti (cuts a dendrogram and optimizes the partition)
  [Return to List of Steps]
Parameters 
  nxpar:   maximum number of requested dendrogram cut.
  niter:   number of "consolidation" iterations (0=no).
  ledin:   printing the correspondence classes-individuals
           (0=no;   1=class-->individuals; 
            2=individuals-->classes; 3= both 1 and 2).
  naxed:   number of axes for printed test-values (0=no).


permu (permutating variables(columns)and individuals (rows) )
  [Return to List of Steps]
Parameters 
                and individuals (rows) )
  ledit:   printing the number of permuted elements 
             (0=no, 1=yes).


polex (positionning words on principal spaces)
  [Return to List of Steps]
Parameters 
  ngraf:   number of requested graphics
  npage:   number of pages of the graphics
  nrows:   number of lines of the graphics


posit (positionning categories in textual principal space)
  [Return to List of Steps]
Parameters 
  naxe:    number of printed axes
  ngraf:   number of requested graphics
  npage:   number of pages for the graphics
  nrows:   number of lines for the graphics
  zoom:    maximum distance to the origin 
             (as numbers of standard deviations)


prico (principal components analysis)
  [Return to List of Steps]
Parameters 
  lcorr:  analysis of the correlation matrix 
            (0=no, 1=yes).
  naxe:   number of requested axes
  nvedi:  number of eigenvalues (histogram).
  ledin:  printing the coordinates of the individuals 
            (0=no, 1=act, 2=tot).
  naxed:  number of printed coordinates.
  navir:  number of digits (right to the dot) 
            (printing of the coordinates).
  nsimu:  number of bootstrap replication 
            (0 = no bootstrap)
  nboot:  type of bootstrap replication 
            (0 = "total bootstrap", 1= partial 
              bootstrap, 2 = bootstrap on variables)

recar (characteristic responses)
  [Return to List of Steps]
Parameter 
  norep:   number of characteristic responses
             for each text


recip (hierarchical clustering: reciprocal neighbours)
  [Return to List of Steps]
Parameters 
  naxu:    number of kept axes.
  nterm:   number of kept terminal elements of tree.
  nived:   number of levels in histogram of indices 
             (0=no).
  lterm:   printing the composition of terminal elements 
             (0=no, 1=yes).
  naxed:   number of coordinates for printing terminal 
              elements  (0=no).
  ldesc:   describing the nodes of the tree 
             (0=no, 1=yes).
  ldend:   printing the dendrogram 
             (0=no, 1=dense, 2=large).


regre (regression and analysis of variance)
  [Return to List of Steps]
Parameters 
  lsupr:   missing value 
              ( =0 : individuals  with m.v. are kept)
              ( =1 : individuals  with m.v. discarded)
              ( default = 1)
  lstat:   printing statistics (=1)  or not (default=0)
  lmat:    printing covariance matrix 
               (default = 0 = no)
               (1 = covariance, 2 = correlation)
      Then: regression formulae: e.g.:   
       V8 = V9 + V57 (one per line)  
       (V8, V9, V57 = continuous variables)
       V8 = V67 + V12*V15  
       (V8 = continuous, V67, V12, V15 = categorical)
      Ending the commands with the key word  "end" 


segme (computing repeated segments)
  [Return to List of Steps]
Parameters 
  nxlon:   maximum length of the segment
  nxseg:   maximum number of requested segments
  nseug:   frequency threshold of segments( general)
  nseu2:   frequency threshold of segments of length 2
  nseu3:   frequency threshold of segments of length 3
  nxle2:   maximum number of segments of length 2
 

selec (selecting active and illustrative variables and individuals)
  [Return to List of Steps]
Parameters 
  lseli:   mode of selection of individuals 
            (0=all, 1=list, 2=filt).
  imass:   index of variable 'weight of individuals' 
              (0=uniform weights).
  lzero:   coding missing responses (0=norec, 1=rec).
  ledit:   printing dictionary of selected variables 
              (0=no, 1=short,2=long).
  nxfil:   maximum number of constants used in filter 
              (lseli = filt ou 2).
  The selection commands use following key words :
            act  ill  end  nomi  cont  freq 


selox (selecting an open question)
  [Return to List of Steps]
Parameters 
  numq:    index of the selected question
            if numq = -1 or numq = list : several questions 
            will be merged (list of questions numbers
            must follow next line)
  lseli:   mode of selection of the individuals 
            (0=tot, 1=list, 2=filt).
  nxfil:   maximum number of constantes used in filter 
            (lseli = filt or 2).
  ldona:  presence of files ndona and ndica
           (files produced by step ardat)
 

semis (partitioning through k-means method)
  [Return to List of Steps]
Parameters 
  naxu:    number of kept axes.
  niter:   number of iterations for each basic partition.
  lmeth:   type of clustering (1=mobil,2=stabl).
  seuil:   number of kept terminal elements
  lresi:   building of a resisual cluster (1=yes,0=no).
  nived:   number of levels in the histogram of indices 
            (0=no).
  ldesc:   describing nodes of the tree (0=no, 1=yes).
  ldend:   printing dendrogram (0=no, 1=dense, 2=large).


setex (setting a new threshold to select words)
  [Return to List of Steps]
Parameters 
  nseu:    threshold of frequency for selecting words.
  nmomi:   minimum number of letters of a kept word.
  nremi:   minimum number of words of a kept response.
  nxmsu:   maximum number of words explicitly eliminated.
  ledit:   printing the dictionaries    
            (0=no, 1=new, 2=tot).


splum (selecting text categories in CA of lexical table)
  [Return to List of Steps]
Parameters 
  ledit:   printing identifiers of supplementary texts.
             --- key-word of headings
                  forme text end


stats (basic statistics for numerical and nominal variables)
  [Return to List of Steps]
Parameters 
  lcont:  printing elementary statistics about numerical
            variables (0=no, 1=yes).
  lhist:  printing the histograms of numerical variables 
           (0=no, 1=yes).
  unit:   default unit for histograms.
  zmin:   minimal default lower bound for histograms.
  zmax:   maximal default lower bound for histograms.
  nclas:  default number of classes for histograms.
  lcom:   commands for numerical variables (0=no, 1=yes).
  nxvdi:   number maximum of values for a discretized 
           numerical variable.
   --- key-words of commands
   zmin zmax unit nclas  hist  libre disc  end


table (cross tabulations with filters)
  [Return to List of Steps]
Parameters 
  imass:   index of the variable 'weight of individuals' 
            (0=uniform weight generated = default).
  lzero:  coding missing responses 
            (0= norec = missing values ignored - default).
            (1 = rec = m.v. become a new category)
  lfilt:  use of a filter (0 = default = no filter)
            (1 = a filter will follow)
  ledic:  printing variables dictionary
            (0 = default = no, 1 = short, 2 = large)
  letyp:  printing standard deviation 
            (case of a table of mean values)
            (0 = no = default)
            (1 = yes: means and std dev. are printed)
  lnull:  printing columns whose weight is zero
            (0 = no, 1 = yes = default)
 

talex (stacked lexical tables)
  [Return to List of Steps]
Parameter 
  ledit:   printing the table words*texts (0=no, 1=yes).