Process Help

DISCAN - multiple Discriminant Analysis

 

Process Name

Menu Path

Link to Command Table

DISCAN

Sample Analysis ribbon | Geochemical Processes | Multiple Discriminant Analysis

Click here

 

Introduction

The calculation of discriminant functions (Discriminant analysis) will establish a set of rules using geological control groups or training sets which can then be applied to the classification of unknown samples using the DISCLA process.

How to use

As a check, the rules as defined by the discriminant functions and group centroids can then be used to reclassify the training sets to see how many mismatches occur.

The linear discriminant function can be considered as a planer separating boundary between the different groups as classified by their input variables (fields), eg. CU, PB and ZN. The group centroids are the fixed positions in space locating the centres of each group cluster as defined by the variables.

File Handling

The input file (&IN) must contain a sample identifier field (@SAMPID) and a group identifier field (@GROUPID). Data from the different control groups must have a unique identifier common to all samples within the group. All the different control groups must be put into one file using the APPEND process before using DISCAN. The group identifier will be representative of a geological control group, eg. to identifier them as granite, mineralised zone, background, lithology A or lithology B etc..

There are optional output files for the calculated discriminant functions (&FUNCTS), group centroids (&CENTRDS) and discriminant scores *&(SCORES). If the multiple discriminant functions and group centroids are required to classify unknown data in the DISCLA process then the relevant output files must be named. If an output file of the re-classified control data is required then the file &SCORES must be used.

If the user wishes to plot maps of the output scores then the scores file can be joined to the original input file using the JOIN process and defining SAMPID as the keyfield.

Files, Fields and Parameters

Input Files

Name

Description

I/O Status

Required

Type

IN

Input file. Must contain group and sample identity fields. A maximum of ten groups are evaluated. The file must be sorted on the *GROUPID field.

Input

Yes

Undefined

Output Files

Name

I/O Status

Required

Type

Description

CENTRDS

Output

No

Undefined

Optional output file containing group centroids identified by the field named in GROUPID.

FUNCTS

Output

No

Undefined

Optional output file containing discriminant functions.

SCORES

Output

No

Undefined

Optional output file containing the calculated scores for the individual samples and a misclassification field.

Fields

Name

Description

Source

Required

Type

Default

GROUPID

Compulsory group identifier field contained in input file IN.

IN

Yes

Alphanumeric

Undefined

SAMPID

Compulsory sample identifier field contained in input file IN.

IN

Yes

Any

Undefined

F1

First field to be used. No fields specified means all.

IN

No

Numeric

Undefined

F2

Second field to be used.

IN

No

Numeric

Undefined

F3

Third field to be used.

IN

No

Numeric

Undefined

F4

Fourth field to be used.

IN

No

Numeric

Undefined

F5

Fifth field to be used.

IN

No

Numeric

Undefined

F6

Sixth field to be used.

IN

No

Numeric

Undefined

F7

Seventh field to be used.

IN

No

Numeric

Undefined

F8

Eighth field to be used.

IN

No

Numeric

Undefined

F9

Ninth field to be used.

IN

No

Numeric

Undefined

F10

Tenth field to be used.

IN

No

Numeric

Undefined

Parameters

Name

Description

Required

Default

Range

Values

RESUM

 

Option

Description

(0)

Do not print summary statistics.

1

Do print summary statistics.

No

0

0,1

0,1

PRIMAT

 

Option

Description

(0)

Option to display sums of squares matrices. Set to 1 for matrices to be displayed.

No

0

0,1

0,1

PRISCO

 

Option

Description

(0)

Option to display group scores. Set to 1 for scores to be displayed. Note: do not use for large files.

No

0

0,1

0,1

Notes

There is a maximum limit of 10 control groups and 45 variables. All individual control groups must also be appended into one file prior to analysis.

Example

!DISCAN

&IN(GG), &CENTRDS(GGCENT), &FUNCTS(GGFUNCTS), @GROUPID='GROUPID', @SAMPID= 'ID', @RESUM=1, @PRIMAT=1

Error and Warning Messages

Message

Description

Solution

*** Warning *** GROUPID field has nn words. Only the first 3 will be used.

 

 

 

 

 

*** Warning *** SAMPID field has nn words. Only the first 3 will be used

 

 

 

 

 

*** Error *** GROUPID field (fieldname) is not in input file.

 

 

 

 

 

*** Error *** No numeric fields on file (filename)
>>> ERR 122 <<< ( fileno.) IN DISCAN

 

 

 

 

 

*** Error *** SAMPID field (fieldname) is not in input file.

 

 

 

 

 
 

*** Error *** in reading data from file (filename)
>>> ERR 200 <<< ( fileno.) IN DISCAN

 

 

 

 

 

*** Error *** There is only one group in file (filename)
Discriminant analysis process terminated.

 

 

 

 

 

*** Error *** There is only one sample in each group
Discriminant analysis process terminated.

 

 

 

 

 

*** Error *** Compulsory file, field or parameter missing
>>> ERR 120 <<< ( fileno.) IN DISCAN