SURVO MM Help System (web edition)

The output file which stores results in text form is not always
sufficient for numerical results obtained in statistical analyses, for
example. To preserve complete numerical accuracy and to provide
immediate access to the matrix interpreter (MAT and MATRUN operations),
the statistical operations will save their output matrices in matrix
files, too.
Such output matrices are always saved on the current data path using
standard names with extension .M (instead of .MAT which is the default
extension for MAT files). They are, however, normal matrix files which
can be processed by MAT operations.
The default names used for output matrices are described below and the
user can change them easily by the OS command REN. For example,
>REN A:*.M *.MAT changes on disk A: all matrix result files (with
extension .M) to matrix files with extension (.MAT).
To see the current .M files, activate, for example
DD *.M

Typical matrix files for statistical results:

CORR.M      Correlation matrix (by operation CORR, for example)
MSN.M       Matrix of 3 columns (by CORR):
              column 1: means of variables
              column 2: standard deviations of variables
              column 3: number of observations

REG.M       Column vector of regression coefficients (LINREG,REGDIAG)
REGS.M      Regression coefficients and their std.errors (LINREG,REGDIAG)
REG_CORR.M  Correlations of regression coefficients (LINREG,REGDIAG)
LG.M        Scalar results of regression analysis (LINREG)
RG.M        Scalar results of regression analysis (REGDIAG)

EST_PAR.M   parameter estimates and their standard errors (ESTIMATE)
EST_CORR.M  correlations of parameter estimates (ESTIMATE)
EST.M       scalar results (ESTIMATE)

STAT.M      certain results of STAT
STAT.M      results of MINSTAT

Matrix files (continued)

PCOMP.M     Principal component loadings (/PCOMPR, /PCOMPCOV)
PCOMPV.M    Variances of components (/PCOMPR, /PCOMPCOV)
PCOEFF.M    Coefficients of pcomp scores (/PCOMPR, /PCOMPCOV)
PCOCENT.M   Variances of components (%) (/PCOMPR, /PCOMPCOV)
PCOEFF.M    Coefficients for principal component scores (/PCOMPR, /PCOMPCOV)

FACT.M      Factor matrix (FACTA)
PFACT.M     Factor matrix (MATRUN PFACT)
AFACT.M     Rotated factor matrix (ROTATE)
TFACT.M     Rotation matrix (ROTATE)
RFACT.M     Factor correlation matrix (ROTATE)
SFACT.M     Factor structure matrix (ROTATE)
FCOEFF.M    Factor score coefficients (/FCOEFF, /FTCOEFF)

L.M         Transformation matrix (/TRAN-SYMMETR, /TRAN-LEASTSQR)
E.M         Residual matrix (/TRAN-SYMMETR, /TRAN-LEASTSQR)

Matrix files (continued)

XTAB.M      Design matrix (TABFIT)
PCOV.M      Covariance estimates of parameters (TABFIT, GENREG)
GENREG.M    Parameter estimates (GENREG)

DISCRL.M    Discriminant coefficients (/DISCRI)
DISCRXR.M   Correlations between variables and discriminators (/DISCRI)

LCAN.M      Vector of canonical correlations (CANON, /CANCORR)
XCAN.M      Correlations of canonical variables with X variables (CANON)
YCAN.M      Correlations of canonical variables with Y variables (CANON)
XCOEFF.M    Coefficients for the first set of canonical variables (CANON)
YCOEFF.M    Coefficients for the second set of canonical variables (CANON)

Matrix files (continued)

CR_COORD.M  Column coordinates (CORRESP)
CR_SUPPC.M  Supplementary column coordinates (CORRESP)
CR_V.M      V from SVD (CORRESP)
CR_CONTR.M  Columnwise absolute contributions (CORRESP)
CR_CORR2.M  Columnwise squared correlations (CORRESP)
CR_MDIST.M  Mass (1st column) and chi^2 distance (2nd column) (CORRESP)

SIMPLEX.M   Solution vector (SIMPLEX)
SLACK.M     Values of the slack variables (SIMPLEX)
TSIMPLEX.M  Simplex Output Table (SIMPLEX)

CSCAL.M     Scale values (/CSCAL)
CSEIGEN.M   Eigenvalues (/CSCAL)
CSCENT.M    Eigenvalues in percentages (/CSCAL)
CSDIST.M    Reproduced distances (/CSCAL)

Matrix files (continued)

LSCAL.M     Estimated configuration matrix (LSCAL)
LSDIST.M    Reproduced distances (LSCAL)

T2COEFF.M   Coefficients of the linear discriminant function (/T2)

CHI2FM.M    Original table with margins (/CHI2)
CHI2E.M     Expected frequencies (/CHI2)
CHI2C.M     Contributions to Chi-square (/CHI2)
CHI2.M      Chi-square value (/CHI2)
CHI2DF.M    Degrees of freedom (/CHI2)

RCOV.M      Residual covariance matrix (RELIAB by K.Vehkalahti)
RCORR.M     Residual correlation matrix (RELIAB)
WEIGHT2.M   Second order scales (RELIAB)

Matrix files (continued)

OR.M        Odds Ratios and confidence intervals (/LOGREG by K.Vehkalahti)

COV.M       Covariance matrix from CORR.M and MSN.M (/COV)

L1.M        Optimal transformation matrix (TRAN1)

MNCOEFF.M   Coefficient matrix (MNSIMUL)
MNMEAN.M    Vector of means (MNSIMUL)

TAILFREQ.M  Relative frequencies of values exceeding treshold values (STATMSF)

COVVAR.M    Optimally permuted covariance matrix (MULTVAR)

Matrix files (continued)

CORR.M      Correlation matrix (CORRMV)
MSN.M       Matrix of 3 columns (see CORR above) (CORRMV)
PAIRFREQ.M  Frequencies of observations for each pair of observations (CORRMV)

BURT.M      Burt's table (BURT by K.Vehkalahti)
CLASS.M     Vector of the number of classes per classifier (BURT)

SMOOTH.M    Smoothened values (LOWESS by J.Puranen)

  M = More information on multivariate statistical operations 
  L = Linear combinations of variables (LINCO) 
  S = More information on statistical operations 
  C = More information on output control 

More information on Survo from
Copyright © Survo Systems 2001-2012.