Measures of central tendency - mode

Thanks to Andy Liaw and J.R. Lockwood for your suggestions.  The which.max() worked great along with the rownames.  The complete solution for me was:

a <- table(varname)
my.mode <- rownames(a)[which.max(a)]
my.mode
[1] "1"

Amazing how a simple concept such as mode can present problems for us.  Thanks again.

To reply to Spencer Graves' question, I didn't find the disucssion via search.  I guess I might have overlooked the thread if it were titled kernel density since that seemed far too technical for this basic topic.  Sorry if I cluttered up the list though.

Cheers,

Patrick

On Fri, 30 Jan 2004 16:47:54 -0500 (EST)

it is an annoyance that table() provides the values being tables as
the rownames of the resultant vector.  you can do something like:

a<-table(x)
rownames(a)[which.max(a)]

On Fri, 30 Jan 2004, Patrick E. McKnight wrote:

Date: Fri, 30 Jan 2004 13:55:19 -0700
From: Patrick E. McKnight <pem at theriver.com>
To: "r-help at stat.math.ethz.ch" <r-help at stat.math.ethz.ch>
Subject: [R] Measures of central tendency - mode

Greetings,

This seems too rudimentary to ask but for the life of me I cannot locate a readily easy method to compute the univariate mode.  I know "mode" is not correct and "table" provides a reasonable count but I figured there would be an easy way to extract the value from the table after I do something like:

max(table(mydadat$myvar))

unfortunately it only returns the max count and not the value that is observed most.  Would some kind soul help me out with this seemingly trivial problem?

Thanks in advance.

Cheers,

Patrick

______________________________________________
R-help at stat.math.ethz.ch mailing list
https://www.stat.math.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html

J.R. Lockwood
412-683-2300 x4941
lockwood at rand.org
http://www.rand.org/methodology/stat/members/lockwood/

Evidently, I didn't read your question carefully enough.  If you 
want the mode of continuous data, that is not well defined, though there 
are devices to estimate such assuming, e.g., a specific distribution or 
a general unimodal distribution or ... .  This was discussed last Dec. 
12-13  by Ted Harding, Brian Ripley and others.  If you are interested, 
you can go www.r-project.org -> search -> "R site search" -> "harding 
mode".  When I did this just now, the first hit was an email on how to 
find the mode using a kernel density estimator.  Clicking "next in 
thread" a couple of times led me to a comment by Brian Ripley with a 
pointer to a document discussing this. 

      ... in case you are interested in more than what you already have. 

      spencer graves