# [R] lda: scaling to 'disctiminant function'

From: Worik Turei stanton <turwa720_at_student.otago.ac.nz>
Date: Mon 25 Jul 2005 - 11:03:06 EST

Friends

Briefly...

In the documentation for lda in MASS it describes the value 'scaling' as 'a matrix which transforms observations into discriminint functions...'.

How?

Verbosely...

I have a matrix of data. 9 independent variables and describing 3-classes. About 100 observations in total. A 10x100 matrix of data.

I am trying to generate two discriminant functions and i think lda is the R-function I want. I have done this in SPSS using the code (is it 'S' code?)...

"DISCRIMINANT /GROUPS=group(1 3) /VARIABLES=agr min man ps con ser fin sps tc /ANALYSIS ALL /PRIORS EQUAL /STATISTICS=MEAN STDDEV UNIVF COEFF RAW COV TABLE /PLOT=COMBINED MAP /CLASSIFY=NONMISSING POOLED " In R I express this as...
library("MASS") # For lda
library("foreign")
Ind <- cbind(EW\$AGR, EW\$MIN, EW\$MAN, EW\$PS, EW\$CON, EW\$SER, EW\$FIN, EW\$SPS, EW\$TC)
Dep <- EW\$GROUP
LDA <- lda(Dep ~ Ind, prior=c(1,1,1)/3)

So far so good.

I have....

> LDA\$scaling

LD1 LD2

```Ind1 1.3415788 3.689959
Ind2 1.5118460 5.215123
Ind3 1.4480197 3.653298
Ind4 2.1898063 3.875867
Ind5 1.2055849 3.913750
Ind6 0.8257858 3.718032
Ind7 1.3481728 3.699259
Ind8 1.2364320 3.871438
Ind9 2.0652630 2.891655

```

What do I do with this to convert it to the coefficients for the discrimination functions?

cheers
Worik

R-help@stat.math.ethz.ch mailing list