Density

-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-
project.org] On Behalf Of li li
Sent: Wednesday, August 08, 2012 9:03 PM
To: r-help
Subject: [R] Density

Dear all,
   Given a set of observations X, I want to evaluate the kernel density
estimator
at these observed values. If I do the following, I do not get the those
estimated values directly.
Can anyone familiar with this give an idea on how to find out the
estimated
density values for X?

X <- rnorm(100)

density(X)-> den0

den0

Call:

density.default(x = X)


Data: X (100 obs.); Bandwidth 'bw' = 0.354


       x                 y

 Min.   :-3.2254   Min.   :0.0002658

 1st Qu.:-1.6988   1st Qu.:0.0359114

 Median :-0.1721   Median :0.1438772

 Mean   :-0.1721   Mean   :0.1635887

 3rd Qu.: 1.3545   3rd Qu.:0.2866889

 Max.   : 2.8812   Max.   :0.3776935


I did write the code for the kernel density

estimator myself. So once I find a proper bandwidth,

I can use the following function. However, it would be nicer

if there is a more direct way.

fhat <- function(x, X){

+          h <- density(X, bw="SJ")$bw

+          n <- length(X)

+          1/(n*h)*sum(dnorm((x-X)/h))}

est <- numeric(length(X))

for (i in 1:length(X)){est[i] <- fhat(x=X[i], X=X)}

est

Thanks in advance.

      Hannah

	[[alternative HTML version deleted]]