Retaining from standard adjustments

Consider the following scenario: Suppose that the base has been adjusted by the base . Occasionally we will be concerned with partial adjustments, and to this purpose for the partial adjustment results we suppose that we adjust initially by the base , and then partially by base ; and that represents the complete adjustment by the base . Suppose that the notional sample size is potential observations on each element of .

Bearings

 

Syntax

1. BD>keep : b =A

2. BD>keep : -b 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the bearing during an adjustment. The bearing is , a linear combination of the elements of B. This linear combination is retained as an assignment, using the supplied name ``A'', which may not include the character ``.''. The constant part of the adjusted expectation is included in the stored assignment.

The switch remains on until switched off. Consequently, care must be taken not to overwrite a bearing stored as an assignment by ones emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the bearing as an assignment. Any assignment so defined during previous adjustments remain defined.  

Partial bearings

 

Syntax

1. BD>keep : pb =A

2. BD>keep : -pb 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the partial bearing during an adjustment. Assuming the adjustment to be of B by D and then F, the partial bearing is , a linear combination of the elements of B. This linear combination is retained as an assignment, using the supplied name ``A'', which may not include the character ``.''. The constant part of the adjusted expectation is included in the stored assignment.

The switch remains on until switched off. Consequently, care must be taken not to overwrite a bearing stored as an assignment by ones emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the bearing as an assignment. Any assignment so defined during previous adjustments remain defined.  

Discrepancy vectors

 

Syntax

1. BD>keep : dv =A

2. BD>keep : -dv 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the discrepancy vector during an adjustment. The discrepancy vector is B, a linear combination of the elements of B. This linear combination is retained as an assignment, using the supplied name ``A'', which may not include the character ``.''. The constant part of the adjusted expectation is included in the stored assignment.

The switch remains on until switched off. Consequently, care must be taken not to overwrite a discrepancy vector stored as an assignment by ones emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the discrepancy vector as an assignment. Any assignment so defined during previous adjustments remain defined.  

Partial discrepancy vectors

 

Syntax

1. BD>keep : pdv =A

2. BD>keep : -pdv 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the partial discrepancy vector during an adjustment. Assuming the adjustment to be of B by D and then F, the partial discrepancy vector is B, a linear combination of the elements of B. This linear combination is retained as an assignment, using the supplied name ``A'', which may not include the character ``.''. The constant part of the adjusted expectation is included in the stored assignment.

The switch remains on until switched off. Consequently, care must be taken not to overwrite a discrepancy vector stored as an assignment by ones emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the discrepancy vector as an assignment. Any assignment so defined during previous adjustments remain defined.  

Canonical directions

 

Syntax

1. BD>keep : cd =A

2. BD>keep : -cd 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of canonical directions during an adjustment. Suppose that an adjustment results in a resolution transform whose canonical directions are . We restrict attention to canonical directions corresponding to positive canonical resolutions. The number of these, , is available via the operand rmrank  discussed in §21.3. Each canonical direction is a linear combination of the component elements of the base B being adjusted, . These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. will be stored as the assignment whose name is A1, and so forth, being stored as the assignment whose name is ``A ''. The corresponding canonical resolutions are available as the [B/D] operators cr (1) ... cr ( ). The constant part of the canonical direction is included in the stored assignment. The name ``A'' supplied must not include a ``.'' character.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the canonical directions that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the canonical directions as assignments. Any assignments so defined during previous adjustments remain defined.  

Adjusted expectations

 

Syntax

1. BD>keep : e =A

2. BD>keep : -e 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of adjusted expectations during an adjustment. Suppose that the component elements of the base B being adjusted are . Then the adjusted expectation for is , a linear combination of the component elements of the adjusting base D. These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. will be stored as the assignment whose name is A1, and so forth, being stored as the assignment whose name is ``A9''. The constant part of the adjusted expectation is included in the stored assignment. The name ``A'' supplied must not include a ``.'' character.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the adjusted expectations that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the adjusted expectations as assignments. Any assignments so defined during previous adjustments remain defined.

The storing of adjusted expectations as assignments is not available for iterative adjustments.  

Name of the adjusting base

 

Syntax

1. BD>keep : fitted =N

2. BD>keep : -fitted 

where N is any valid name, intended to become the name of a base.

The first form of the syntax is used to retain, using the name supplied, the currently fitted base. Hence, if the current adjustment can be referred to as , then N will become the base consisting of the bases and elements . This command has immediate effect if there exists a current adjustment, so that one need not not wait for a forthcoming adjust-type command. Every new adjustment causes the redefinition of this base, and the base name N used can be such that a former base of the same name is overwritten. The adjusted  argument to the KEEP:  command is similar, but applies to the base being adjusted. The control applies only to the ADJUST:  command.

The second form of the command discontinues retention of the currently fitted base.  

Name of the adjusted base

 

Syntax

1. BD>keep : adjusted =N

2. BD>keep : -adjusted 

where N is any valid name, intended to become the name of a base.

The first form of the syntax is used to retain, using the name supplied, the currently adjusted base. Hence, if the current adjustment can be referred to as [B/D], then N will become the base consisting of the base B. This command has immediate effect if there exists a current adjustment, so that one need not not wait for a forthcoming adjust-type command. Every new adjustment causes the redefinition of this base, and the base name N used can be such that a former base of the same name is overwritten. The fitted  argument to the KEEP:  command is similar, but applies to the base used for adjusting. The control applies only to the ADJUST:  command.

The second form of the command discontinues retention of the currently adjusted base.  

Partial canonical directions

 

Syntax

1. BD>keep : pcd =A

2. BD>keep : -pcd 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of partial canonical directions during an adjustment. Suppose that a partial adjustment results in a partial resolution transform whose partial canonical directions are . We restrict attention to partial canonical directions corresponding to positive partial canonical resolutions. The number of these, , is available via the operand prmrank  discussed in §21.3. Each partial canonical direction is a linear combination of the component elements of the base B being adjusted, . These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. will be stored as the assignment whose name is A1, and so forth, being stored as the assignment whose name is ``A ''. The corresponding partial canonical resolutions are available as the [B/D] operators pcr (1) ... pcr ( ). The constant part of the canonical direction is included in the stored assignment. The name ``A'' supplied must not include a ``.'' character.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the partial canonical directions that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the partial canonical directions as assignments. Any assignments so defined during previous adjustments remain defined.  

Resolution matrix columns

 

Syntax

1. BD>keep : rm =A

2. BD>keep : -rm 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the columns of the resolution matrix, , as assignments during an adjustment. Each column is interpreted as a linear combination of the component elements of the base B being adjusted, . These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. The first column will be stored as the assignment whose name is A1, and so forth. No constant part is included in the stored assignment. The name ``A'' supplied must not include a ``.'' character.

The coefficients making up the resolution matrix are available interactively by using the ascf  function. These numbers reproduce the output available under the rm  option.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the columns that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the resolution matrix columns as assignments. Any assignments so defined during previous adjustments remain defined.  

Maximal resolution matrix columns

 

Syntax

1. BD>keep : mrm =A

2. BD>keep : -mrm 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the columns of the maximal resolution matrix, , as assignments during an adjustment. Each column is interpreted as a linear combination of the component elements of the base B being adjusted, . These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. The first column will be stored as the assignment whose name is A1, and so forth. No constant part is included in the stored assignment. The name ``A'' supplied must not include a ``.'' character.

The coefficients making up the maximal resolution matrix are available interactively by using the ascf  function. These numbers reproduce the output available under the mrm  option.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the columns that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the maximal resolution matrix columns as assignments. Any assignments so defined during previous adjustments remain defined.  

Partial resolution matrix columns

 

Syntax

1. BD>keep : prm =A

2. BD>keep : -prm 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of the columns of the partial resolution matrix,

as assignments during an adjustment. Each column is interpreted as a linear combination of the component elements of the base B being adjusted, . These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. The first column will be stored as the assignment whose name is A1, and so forth. No constant part is included in the stored assignment. The name ``A'' supplied must not include a ``.'' character.

The coefficients making up the partial resolution matrix are available interactively by using the ascf  function. These numbers reproduce the output available under the prm  option.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the columns that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the partial resolution matrix columns as assignments. Any assignments so defined during previous adjustments remain defined.  

Adjusted versions

 

Syntax

1. BD>keep : av =A

2. BD>keep : -av 

where A is the name of an assignment.

The first form of the syntax is used to switch on the retention of adjusted versions during an adjustment. Corresponding to each element is an adjusted version . Each adjusted version is a linear combination of the component elements of the base D being used to adjust B.

These linear combinations are retained as assignments, using the supplied name ``A'' as a stem. Thus, will be stored as the assignment whose name is A1, and so forth. The constant part of every adjusted version is zero. The name ``A'' supplied must not include a ``.'' character.

Be warned that there are subtleties involved in building adjusted versions retained from general or pure exchangeable adjustments; the direct building of the assignments won't produce elements with the beliefs that you might have intended because the assignments ``forget'' about the exchangeable structure of the information source elements.

The switch remains on until switched off. Consequently, care must be taken not to overwrite the adjusted versions that were stored as assignments by those emanating from fresh adjustments, unless this is what is desired.

The second form of the syntax is used to switch off the automatic storing of the adsjusted versions as assignments. Any assignments so defined during previous adjustments remain defined.