NONMEM output

In this chapter we interpret in detail the output of NONMEM runs. We use the following model as an running example. In the previous chapter we have covered all the control records in this model.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
$PROBLEM PHENOBARB WITH WEIGHT IN MODELS FOR CL AND V
$INPUT   ID TIME AMT WT APGR DV
$DATA    INDATA
$SUBROUTINE ADVAN1
$PK
   TVCL=THETA(1)+THETA(3)*WT
   CL=TVCL+ETA(1)
   TVVD=THETA(2)+THETA(4)*WT
   V=TVVD+ETA(2)
                       ; THE FOLLOWING ARE REQUIRED BY PREDPP
   K=CL/V
   S1=V
$ERROR
   Y=F+ERR(1)
$THETA  (0,.0027) (0,.70) .0018  .5
$OMEGA  .000007, .3
$SIGMA   8
$ESTIMATION PRINT=10
$COVARIANCE UNCONDITIONAL
$TABLE ID TIME AMT WT APGR DV

Model input

The first page of NONMEM’s output is shown below. On this page, NONMEM repeats the instructions it was given in the control file and describes the data file. The first page of the output should be checked carefully. Problems in a NONMEM run can often be traced to errors in the problem specification. For example, always check that the initial parameter estimates were entered correctly

 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
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
NONLINEAR MIXED EFFECTS MODEL PROGRAM (NONMEM) VERSION 7.5.2
ORIGINALLY DEVELOPED BY STUART BEAL, LEWIS SHEINER, AND ALISON BOECKMANN
CURRENT DEVELOPERS ARE ROBERT BAUER, ICON DEVELOPMENT SOLUTIONS,
AND ALISON BOECKMANN. IMPLEMENTATION, EFFICIENCY, AND STANDARDIZATION
PERFORMED BY NOUS INFOSYSTEMS.

PROBLEM NO.:         1
PHENOBARB WITH WEIGHT IN MODELS FOR CL AND V
DATA CHECKOUT RUN:              NO
DATA SET LOCATED ON UNIT NO.:    2
THIS UNIT TO BE REWOUND:        NO
CREATE/ADD TO FDATA.csv:        YES
NO. OF DATA RECS IN DATA SET:      744
NO. OF DATA ITEMS IN DATA SET:   8
ID DATA ITEM IS DATA ITEM NO.:   1
DEP VARIABLE IS DATA ITEM NO.:   6
MDV DATA ITEM IS DATA ITEM NO.:  8
INDICES PASSED TO SUBROUTINE PRED:
  7   2   3   0   0   0   0   0   0   0   0
LABELS FOR DATA ITEMS:
ID TIME AMT WT APGR DV EVID MDV
FORMAT FOR DATA:
(6E6.0,2F2.0)

TOT. NO. OF OBS RECS:      155
TOT. NO. OF INDIVIDUALS:       59
LENGTH OF THETA:   4
DEFAULT THETA BOUNDARY TEST OMITTED:    NO
OMEGA HAS SIMPLE DIAGONAL FORM WITH DIMENSION:   2
DEFAULT OMEGA BOUNDARY TEST OMITTED:    NO
SIGMA HAS SIMPLE DIAGONAL FORM WITH DIMENSION:   1
DEFAULT SIGMA BOUNDARY TEST OMITTED:    NO
INITIAL ESTIMATE OF THETA:
LOWER BOUND    INITIAL EST    UPPER BOUND
 0.0000E+00     0.2700E-02     0.1000E+07
 0.0000E+00     0.7000E+00     0.1000E+07
-0.1000E+07     0.1800E-02     0.1000E+07
-0.1000E+07     0.5000E+00     0.1000E+07
INITIAL ESTIMATE OF OMEGA:
0.7000E-05
0.0000E+00   0.3000E+00
INITIAL ESTIMATE OF SIGMA:
0.8000E+01
COVARIANCE STEP OMITTED:        NO
EIGENVLS. PRINTED:              NO
SPECIAL COMPUTATION:            NO
COMPRESSED FORMAT:              NO
GRADIENT METHOD USED:     NOSLOW
SIGDIGITS ETAHAT (SIGLO):                  -1
SIGDIGITS GRADIENTS (SIGL):                -1
EXCLUDE COV FOR FOCE (NOFCOV):              NO
Cholesky Transposition of R Matrix (CHOLROFF):0
KNUTHSUMOFF:                                -1
RESUME COV ANALYSIS (RESUME):               NO
SIR SAMPLE SIZE (SIRSAMPLE):
NON-LINEARLY TRANSFORM THETAS DURING COV (THBND): 1
PRECONDTIONING CYCLES (PRECOND):        0
PRECONDTIONING TYPES (PRECONDS):        TOS
FORCED PRECONDTIONING CYCLES (PFCOND):0
PRECONDTIONING TYPE (PRETYPE):        0
FORCED POS. DEFINITE SETTING DURING PRECONDITIONING: (FPOSDEF):0
SIMPLE POS. DEFINITE SETTING: (POSDEF):-1
TABLES STEP OMITTED:    NO
NO. OF TABLES:           1
SEED NUMBER (SEED):    11456
NPDTYPE:    0
INTERPTYPE:    0
RANMETHOD:             3U
MC SAMPLES (ESAMPLE):    300
WRES SQUARE ROOT TYPE (WRESCHOL): EIGENVALUE
-- TABLE   1 --
RECORDS ONLY:    ALL
4 COLUMNS APPENDED:    YES
PRINTED:               YES
FORMAT:                S1PE11.4
IDFORMAT:
LFORMAT:
RFORMAT:
FIXED_EFFECT_ETAS:
USER-CHOSEN ITEMS:
ID TIME AMT WT APGR
SCATTERPLOT STEP OMITTED:    NO
FAMILIES OF SCATTERPLOTS:     2
-- SCATTERPLOT   1 --
UNIT SLOPE LINE:            YES
ITEMS TO BE SCATTERED:    DV    PRED
-- SCATTERPLOT   2 --
UNIT SLOPE LINE:             NO
ITEMS TO BE SCATTERED:    WT    RES

Most lines above are easy to interpret. We make several comments as follows.

  • Line 21 gives the labels for all the data items. The first six labels are specified in $INPUT. The next two (EVID, MDV) are those of two data items added to the data set by NONMEM (NONMEM deduces EVID and MDV from given data items. It also supplies data item PRED (prediction), RES (residual), and WRES (weighted residual)). In the formulation of \(y=f(x, \phi, t)+\epsilon\), ID, TIME, AMT, WT, and APGR are elements of \(x\), DV is \(y\), and PRED is described by $SUBROUTINE and $PK records.
  • Line 27-43 describe the contents of the $THETA, $OMEGA and $SIGMA. First the number of elements in line 27, 29 and 31, then their initial estimates are displayed. In line 35-38, value 0.1000e+07 and -0.1000e+07 are NONMEM’s way of expressing the values \(\infty\) and \(-\infty\), i.e., of describing THETAs that are unbounded on one or both sides. Line 40-41 show the low triangular part of the \(\Omega\) matrix: the diagonal entries are the variances \(\eta\)'s. Since line 29 states that NONMEM understands OMEGA to be diagonal, the off-diagonal element \(\text{cov}(\eta_1,\eta_2)\) is automatically set to zero.

Output below describes the options of the $ESTIMATION.

 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
#TBLN:      1
#METH: First Order

ESTIMATION STEP OMITTED:                 NO
SHRINK INFO WITH EVALUATION (EVALSHRINK) NO
ANALYSIS TYPE:                           POPULATION
NUMBER OF SADDLE POINT RESET ITERATIONS:      0
GRADIENT METHOD USED:               NOSLOW
EPS-ETA INTERACTION:                     NO
NO. OF FUNCT. EVALS. ALLOWED:            360
NO. OF SIG. FIGURES REQUIRED:            3
INTERMEDIATE PRINTOUT:                   YES
ESTIMATE OUTPUT TO MSF:                  NO
IND. OBJ. FUNC. VALUES SORTED:           NO
NUMERICAL DERIVATIVE
      FILE REQUEST (NUMDER):               NONE
MAP (ETAHAT) ESTIMATION METHOD (OPTMAP):   0
ETA HESSIAN EVALUATION METHOD (ETADER):    0
INITIAL ETA FOR MAP ESTIMATION (MCETA):    0
SIGDIGITS FOR MAP ESTIMATION (SIGLO):      100
GRADIENT SIGDIGITS OF
      FIXED EFFECTS PARAMETERS (SIGL):     100
NOPRIOR SETTING (NOPRIOR):                 0
NOCOV SETTING (NOCOV):                     OFF
DERCONT SETTING (DERCONT):                 OFF
FINAL ETA RE-EVALUATION (FNLETA):          1
EXCLUDE NON-INFLUENTIAL (NON-INFL.) ETAS
      IN SHRINKAGE (ETASTYPE):             NO
NON-INFL. ETA CORRECTION (NONINFETA):      0
RAW OUTPUT FILE (FILE): 2.11.ext
EXCLUDE TITLE (NOTITLE):                   NO
EXCLUDE COLUMN LABELS (NOLABEL):           NO
FORMAT FOR ADDITIONAL FILES (FORMAT):      S1PE12.5
PARAMETER ORDER FOR OUTPUTS (ORDER):       TSOL
KNUTHSUMOFF:                               0
INCLUDE LNTWOPI:                           NO
INCLUDE CONSTANT TERM TO PRIOR (PRIORC):   NO
INCLUDE CONSTANT TERM TO OMEGA (ETA) (OLNTWOPI):NO
ADDITIONAL CONVERGENCE TEST (CTYPE=4)?:    NO
EM OR BAYESIAN METHOD USED:                 NONE

NONMEM uses tag "#" to sectionize output files. The "#TBLN" entry on line 1 indicates estimation step is numbered "1". If additional $ESTIMATION record are used their output will be labeld in corresponding "#TBLN"s. Despite we only used the PRINT option, the output shows all the default options values. For example, line 10 shows the maximum number of evaluations of the objective function during the Estimation Step is 360, and line 11 shows the desired number of significant digits in the final parameter estimate is 3.

NONMEM's built-in ADVAN models, TRANS parameterizations, and $PK block specifications are in fact from a sub-library called PREDPP (PRED of Population Pharmacokinetics). The output below shows the ADVAN and TRANS routines used

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
DOUBLE PRECISION PREDPP VERSION 7.5.2

ONE COMPARTMENT MODEL (ADVAN1)
MAXIMUM NO. OF BASIC PK PARAMETERS:   2
BASIC PK PARAMETERS (AFTER TRANSLATION):
  ELIMINATION RATE (K) IS BASIC PK PARAMETER NO.:  1

COMPARTMENT ATTRIBUTES
COMPT. NO.   FUNCTION   INITIAL    ON/OFF      DOSE      DEFAULT    DEFAULT
                        STATUS     ALLOWED    ALLOWED    FOR DOSE   FOR OBS.
   1         CENTRAL      ON         NO         YES        YES        YES
   2         OUTPUT       OFF        YES        NO         NO         NO

and the parameterization details.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
ADDITIONAL PK PARAMETERS - ASSIGNMENT OF ROWS IN GG
COMPT. NO.                             INDICES
             SCALE      BIOAVAIL.   ZERO-ORDER  ZERO-ORDER  ABSORB
                        FRACTION    RATE        DURATION    LAG
   1            3           *           *           *           *
   2            *           -           -           -           -
            - PARAMETER IS NOT ALLOWED FOR THIS MODEL
            * PARAMETER IS NOT SUPPLIED BY PK SUBROUTINE;
              WILL DEFAULT TO ONE IF APPLICABLE
DATA ITEM INDICES USED BY PRED ARE:
  EVENT ID DATA ITEM IS DATA ITEM NO.:      7
  TIME DATA ITEM IS DATA ITEM NO.:          2
  DOSE AMOUNT DATA ITEM IS DATA ITEM NO.:   3

PK SUBROUTINE CALLED WITH EVERY EVENT RECORD.
PK SUBROUTINE NOT CALLED AT NONEVENT (ADDITIONAL OR LAGGED) DOSE TIMES.
DURING SIMULATION, ERROR SUBROUTINE CALLED WITH EVERY EVENT RECORD.
OTHERWISE, ERROR SUBROUTINE CALLED ONCE IN THIS PROBLEM.

Several notes regarding the above output.

  • In line 5, we see compartment number 1 (central compartment for ADVAN1) has a scaling parameter ("3" indicates storage location and unimportant here). This is corresponding to "S1" assignment in the model. Positions marked with "*", such as the bioavailability fraction, correspond to additional PK parameters that are allowed by the model but that are not assigned a value by $PK statements. Positions marked with "-" correspond to additional parameters that are not permitted. For instance, dose-related parameters are not allowed for compartment 2 (the output compartment), because this compartment cannot receive doses.
  • Line 10-13 describe the locations of PREDPP data items in the input data record. Note that data item 7, Event ID, was appended. Line 15 shows thats by default $PK statements are evaluated with every event record. Lagged and additional doses are discussed later in this document. They are not used in this example.
  • Line 18 shows that the $ERROR statements describe a simple error model that needs be evaluated only once at the beginning of the problem. Line 17 indicates that, should the Simu- lation Step be implemented, PREDPP will disregard this limitation and evaluate the $ERROR statements with every event record, so that randomly-generated values of intra-individual error can be applied at every observation event. (This example does not involve simulation, but the PK and ERROR routines which implement the $PK and $ERROR statements are capable of supporting all NONMEM tasks, including simulations).

Diagnostic output from the estimation step

The output elow is from the Estimation Step.

 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
MONITORING OF SEARCH:


ITERATION NO.:    0    OBJECTIVE VALUE:   675.657684094370        NO. OF FUNC. EVALS.:   7
CUMULATIVE NO. OF FUNC. EVALS.:        7
NPARAMETR:  2.7000E-03  7.0000E-01  1.8000E-03  5.0000E-01  7.0000E-06  3.0000E-01  8.0000E+00
PARAMETER:  1.0000E-01  1.0000E-01  1.0000E-01  1.0000E-01  1.0000E-01  1.0000E-01  1.0000E-01
GRADIENT:  -3.9976E+01 -8.0006E+01 -4.2941E+02 -1.0001E+03  1.5391E+01  5.2661E+01  9.0187E+00

ITERATION NO.:   10    OBJECTIVE VALUE:   613.258949532521        NO. OF FUNC. EVALS.:  84
CUMULATIVE NO. OF FUNC. EVALS.:       91
NPARAMETR:  7.1965E-04  2.5796E-01  4.3459E-03  8.4505E-01  2.2246E-06  9.9144E-02  8.9710E+00
PARAMETER: -1.2222E+00 -8.9828E-01  2.4144E-01  1.6901E-01 -4.7316E-01 -4.5360E-01  1.5728E-01
GRADIENT:   3.7213E+00  6.3503E+00  5.0514E+01  5.2246E+01  2.6041E+00  4.2624E+00  1.4554E+01

ITERATION NO.:   20    OBJECTIVE VALUE:   609.398192823140        NO. OF FUNC. EVALS.:  81
CUMULATIVE NO. OF FUNC. EVALS.:      172
NPARAMETR:  1.1805E-04  8.8326E-02  4.6808E-03  9.3708E-01  1.3391E-06  7.3110E-02  8.7520E+00
PARAMETER: -3.0299E+00 -1.9700E+00  2.6004E-01  1.8742E-01 -7.2697E-01 -6.0591E-01  1.4492E-01
GRADIENT:   1.6349E-01 -5.8676E-01 -6.1457E+00 -1.4122E+01 -5.9614E-02  3.5384E-01  1.4194E-03

ITERATION NO.:   30    OBJECTIVE VALUE:   609.179165845924        NO. OF FUNC. EVALS.:  99
CUMULATIVE NO. OF FUNC. EVALS.:      271
NPARAMETR:  2.7000E-05  1.1935E-01  4.7519E-03  9.1899E-01  1.3698E-06  7.5079E-02  8.7026E+00
PARAMETER: -4.5116E+00 -1.6690E+00  2.6399E-01  1.8380E-01 -7.1561E-01 -5.9262E-01  1.4209E-01
GRADIENT:   0.0000E+00  8.9319E-03  1.4467E-01  5.4480E-01 -1.7084E-03 -3.2719E-03 -7.7586E-03

ITERATION NO.:   31    OBJECTIVE VALUE:   609.179165845924        NO. OF FUNC. EVALS.:   0
CUMULATIVE NO. OF FUNC. EVALS.:      271
NPARAMETR:  2.7000E-05  1.1935E-01  4.7519E-03  9.1899E-01  1.3698E-06  7.5079E-02  8.7026E+00
PARAMETER: -4.5116E+00 -1.6690E+00  2.6399E-01  1.8380E-01 -7.1561E-01 -5.9262E-01  1.4209E-01
GRADIENT:   0.0000E+00  8.9319E-03  1.4467E-01  5.4480E-01 -1.7084E-03 -3.2719E-03 -7.7586E-03

#TERM:
MINIMIZATION SUCCESSFUL
NO. OF FUNCTION EVALUATIONS USED:      271
NO. OF SIG. DIGITS IN FINAL EST.:  3.7
PARAMETER ESTIMATE IS NEAR ITS BOUNDARY

TOTAL DATA POINTS NORMALLY DISTRIBUTED (N):          155
N*LOG(2PI) CONSTANT TO OBJECTIVE FUNCTION:    284.870945293449
OBJECTIVE FUNCTION VALUE WITHOUT CONSTANT:    609.179165845924
OBJECTIVE FUNCTION VALUE WITH CONSTANT:       894.050111139373
REPORTED OBJECTIVE FUNCTION DOES NOT CONTAIN CONSTANT

TOTAL EFFECTIVE ETAS (NIND*NETA):                           118

#TERE:
Elapsed estimation  time in seconds:     0.61
S MATRIX ALGORITHMICALLY SINGULAR
S MATRIX IS OUTPUT
INVERSE COVARIANCE MATRIX SET TO RS*RMAT, WHERE S* IS A PSEUDO INVERSE OF S
Elapsed covariance  time in seconds:     0.16
Elapsed postprocess time in seconds:     0.10

Lines 1 through 33 are referred to as the intermediate output. (Iteration 0 indicates the initial estimation). . With NONMEM 7, the parameter estimates are displayed in their natural (unscaled) space, as indicated by "NPARAMETR" labels. "PARAMETER" labels give the unconstrained parameter (UCP) estimates. The parameter search is carried out in a different parameter space, before transformed to UCP. In the transformation process a scaling occurs so that the initial estimate of each of the UCP is 0.1. Thus, in line 7, all parameters are .1 at iteration 0. Parameters are output in the order of \(\theta\) elements, \(\Omega\) elements, \(\Sigma\) elements. In this example, they are \(\theta_1\), \(\theta_2\), \(\theta_3\), \(\theta_4\), \(\Omega_{11}\), \(\Omega_{22}\), and \(\Sigma_{11}\). "GRANDENT" labels the partial derivatives of the objective function with respect to the parameter.

As the iterations progress, we see that both the objective function value and parameters change rapidly at first and less so as the solution closes in the final values. In the meantime, the gradients approach 0.

Note that The values computed for the gradients are very sensitive to differences in computer arith- metic and precision. If a given NONMEM run is repeated on a different computer, or on the same computer with different machine precision or a different FORTRAN compiler, it is likely that the gradients will be different. This will cause the search to follow a different path to the minimum, so that lines 4 through 33 may be quite different. However, each final estimate of a parameter should always be the same to the number of requested significant digits. (Minor differences may also be observed in the output of the Covariance Step, below; this output is also sensitive to computational differences).

Lines 34 to 37 are always printed, regardless if the itermediate printout is suppressed or not. The summary output statements of the estimation results begins with the tag "#TERM" on line 34 and ends with the #TERE tag on line 48. Line 35, "MINI- MIZATION SUCCESSFUL", signifies that the search appears to have located a minimum of the objective function. The final parameter estimates must be examined in their (untransformed) space, based on domain knowledge and physiological plausibility, among others. The Estimation Step is not always successful and we will discuss the error messages later.

The number of significant digits in the final estimate is given in line 37 as 3.7. This can be interpreted as meaning that no (transformed) parameter estimate is actually determined to less than 3.7 significant digits. More specifically, when the UCP estimates were compared between the last two iterations, none differed in the first (almost) 4 significant figures including leading zeros after the decimal point. Because NONMEM displays only 3 significant digits in the printed parameter estimates, and for other reasons as well, by default NM-TRAN requests only 3 significant digits.

Covariance step output

Output is from the Covariance Step, requested via the $COVARIANCE record. It shows the standard errors of the parameter estimates. Note that "…….." denotes a value (such as the standard error for \(\hat{\Omega}_{12}\)) that is 0 by definition, rather than due to computation.

 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
************************************************************************************************************************
********************                                                                                ********************
********************                                   FIRST ORDER                                  ********************
********************                            STANDARD ERROR OF ESTIMATE                          ********************
********************                                                                                ********************
************************************************************************************************************************

THETA - VECTOR OF FIXED EFFECTS PARAMETERS   *********
        TH 1      TH 2      TH 3      TH 4
        0.00E+00  1.48E-01  2.24E-04  1.13E-01

OMEGA - COV MATRIX FOR RANDOM EFFECTS - ETAS  ********
        ETA1      ETA2
ETA1    7.30E-07
ETA2   .........  3.63E-02

SIGMA - COV MATRIX FOR RANDOM EFFECTS - EPSILONS  ****
        EPS1
EPS1    1.71E+00

OMEGA - CORR MATRIX FOR RANDOM EFFECTS - ETAS  *******
        ETA1      ETA2
ETA1    3.12E-04
ETA2   .........  6.63E-02

SIGMA - CORR MATRIX FOR RANDOM EFFECTS - EPSILONS  ***
        EPS1
EPS1    2.90E-01

The low triangular output of the covariance matrix, the correlation matrix, and the inverse of covariance matrix are similar and omitted here.

Table output

Below is the first 12 lines of the table file produced by the $TABLE record. The ID indicate it is from the first individual.

 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
TABLE NO.  1



LINE NO.     ID      TIME       AMT        WT      APGR        DV      PRED      RES       WRES

   1    1.00E+00  0.00E+00  2.50E+01  1.40E+00  7.00E+00  0.00E+00  1.78E+01  0.00E+00  0.00E+00

   2    1.00E+00  2.00E+00  0.00E+00  1.40E+00  7.00E+00  1.73E+01  1.76E+01 -3.14E-01 -2.92E-01

   3    1.00E+00  1.25E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  1.92E+01  0.00E+00  0.00E+00

   4    1.00E+00  2.45E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.07E+01  0.00E+00  0.00E+00

   5    1.00E+00  3.70E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.20E+01  0.00E+00  0.00E+00

   6    1.00E+00  4.80E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.33E+01  0.00E+00  0.00E+00

   7    1.00E+00  6.05E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.45E+01  0.00E+00  0.00E+00

   8    1.00E+00  7.25E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.56E+01  0.00E+00  0.00E+00

   9    1.00E+00  8.53E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.66E+01  0.00E+00  0.00E+00

  10    1.00E+00  9.65E+01  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.77E+01  0.00E+00  0.00E+00

  11    1.00E+00  1.08E+02  3.50E+00  1.40E+00  7.00E+00  0.00E+00  2.87E+01  0.00E+00  0.00E+00

  12    1.00E+00  1.12E+02  0.00E+00  1.40E+00  7.00E+00  3.10E+01  2.81E+01  2.88E+00  6.88E-01

Each row in the table corresponds to a record of the input file, and the rows appear in the same order as do the corresponding records of the input data file. Note that the values of RES and WRES are always shown as zero for non-observation records†, whereas a (possibly) nonzero value of PRED is printed for every record.

If there are more than 900 data records, separate tables are produced for groups of 900 records. The last table contains the remaining records. If the rows of the table are sorted, each group of records is sorted separately. When the input data file is large, the table will require many pages to print. Therefore, the $TABLE record should be omitted unless needed for diagnostic purposes (such as when initially checking a new data set or model).