# RE: [R] Using R for classifying new samples

From: bogdan romocea <br44114_at_gmail.com>
Date: Sat 28 May 2005 - 01:18:32 EST

Read this book, "Multivariate Statistical Analysis: A Conceptual Introduction" by Sam Kash Kachigan. I think it's *great*, and perfect for someone without any statistical background.

-----Original Message-----
From: manav ram [mailto:culaterz2003@yahoo.com] Sent: Friday, May 27, 2005 10:56 AM
To: r-help@stat.math.ethz.ch
Subject: [R] Using R for classifying new samples

Hello,
I do not have any statistical background, So I shall apologise if I am asking trivial question or help.

I am trying to work with R.
The problem I have on hand is:
I have 2 sets of data(means & SD for each sample in the group of both sets). The sample size is massive(2000+ in each grp).
I have a new set of experimental data and I like to classify this to either of the grps based on statistical evidence.
I thought using Linear discriminant analysis of R(MASS package) I can solve this problem.
But unfortunately I am not quite sure its the right way and further with given my background I am not even sure how to prepare the data to test.

To explain my problem further, I have prepared a small set of data to be tested....so that it might be easy to suggest the right way to go ahead.The data is in the
form""Object>Query>mean_grp1>SD_grp1>mean_grp2>SD_grp2""

```A>3.890>3.315>1.105>1.395>0.345
B>0.915>1.193>0.334>0.638>0.034
C>2.059>2.155>0.614>1.042>0.159
D>1.372>0.901>0.314>0.384>0.174

```

What I would like to test is if the "query" belongs to grp 1 or 2(for all my samples,2000+)and if it can read out to me teh query belongs to grp_1 or grp_2. Is there some good examples that I can refer to which can teach me step wise right from preparing the data to testing???or would there be someone who can help me with this problem?????

Thanks a lot for your time...look forward to hear some help and suggestions

Manav

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