RE 1 The number of degrees of freedom has not been specified.

RE 2 Model formula contains an operator not valid in regression models.

RE 3 Model formula contains too many variates and factors.

RE 4 Number of binomial trials is 1 and the dispersion parameter is not fixed.

RE 5 The `SAVE`

option of `FSSPM`

has changed during sequential accumulation.

RE 6 There are no units with values present for all vectors and positive weights.

RE 7 Numbers of binomial trials have not been supplied.

RE 8 The distribution is invalid for the type of model chosen.

RE 9 The type of model chosen cannot be analysed within groups.

RE 10 Regression models cannot be fitted until a `MODEL`

statement is given.

RE 11 The number of response variates is invalid for the type of model chosen.

RE 12 The `SSPM`

does not include information on a current response variate.

RE 13 The parameter list was ignored because an `SSPM`

was supplied.

RE 14 An option of `TERMS`

was ignored because an `SSPM`

was supplied.

RE 15 An option of `MODEL`

was overridden because an `SSPM`

was supplied.

RE 16 There is not enough space to form all sums of squares and products.

RE 17 A response variate cannot be used as an explanatory variate as well.

RE 18 The chosen model cannot be based on an `SSPM`

.

RE 19 Iteratively fitted model cannot be based on a sequentially formed `SSPM`

.

RE 20 The `PRINT`

option was ignored because a saved `SSPM`

was supplied.

RE 21 Single precision `SSPM`

values may not be sufficient for accurate regression.

RE 22 Statements to modify regression model must be preceded by a `TERMS`

statement.

RE 23 A term is not in the maximal model defined by the `TERMS`

statement.

RE 24 The status of the constant term cannot be changed in this model.

RE 25 A data value is incompatible with the model.

RE 26 A fitted value predicted by the model is out of valid range for the model.

RE 27 Iterative fitting process has diverged, probably because of extreme data.

RE 28 Iterative weights have become 0, or have been held at a limit.

RE 29 The iterative fitting process has not converged.

RE 30 `STEP`

and `TRY`

statements cannot be used with nonlinear models and curves.

RE 31 There are no degrees of freedom to calculate residual mean squares.

RE 32 Results cannot be displayed or kept because no model has been fitted.

RE 33 `PREDICT`

statements can be used only for linear or generalized linear models.

RE 34 `RFUNCTION`

statements can be used only for nonlinear models or curves.

RE 35 Requested results from the current model are not available.

RE 36 Predictions cannot be formed.

RE 37 Factors to classify predictions do not match those classifying weight table.

RE 38 A structure in the `Y`

parameter is not a response variate.

RE 39 A response variate has all values zero or missing.

RE 40 Invalid submodel specification.

RE 41 Attempt to fit submodel that was not included in maximal model.

RE 42 Invalid calculation for calculated distribution or link.

RE 43 Matrix of weights cannot be used.

RE 44 Stepwise search is alternating between two models.

RE 45 Negative ratio in summary of regression analysis.

RE 46 Failure to carry out local smoothing.

RE 47 Local smoothing trace evaluation outside allowed range.

RE 48 Singularity encountered in calculation of local smooth.

RE 49 No observations found at the reference level of a factor.

RE 50 Regression save structure was formed with a previous release of Genstat.

RE 51 Missing values not allowed in `REG`

matrix.

RE 52 Likelihood-based lower confidence limit failed (set to missing).

RE 53 Likelihood-based upper confidence limits failed (set to missing).

RE 54 Cannot use marginal weights in `PREDICT`

when `MVINCLUDE=explanatory`

in `TERMS`

.

RE 55 Too many columns to fit in `LOESS`

regression (increase smoothing parameter).

RE 56 Unrecognised parallel setting.

RE 57 Incorrect number of levels in parallel setting.

RE 58 Cannot have `MVINCLUDE=explanatory`

, missing `GROUPS`

factor values and higher-order terms with the `GROUPS`

factor.

RE 59 Cannot save estimates for terms involving submodels.

RE 60 Regression quantiles could not be calculated.

RE 61 Predictions may be unreliable as fit was stopped before convergence.

RE 62 Too many parameters to form ssp matrix (limit is 65535).

RE 63 Regression data structure (e.g. in `MODEL`

) has been deleted.

RE 64 Parameter names too long

# Regression analysis fault codes

Updated on September 2, 2019