Can't get a different density function to the histogram distribution plotted

Asked Apr 07 '18 at 23:34

Active Apr 08 '18 at 16:34

Viewed 53 times

I want to test the normality of a statistical exeperience defined by this code.

# a

y0<-50
a<-10
b<-0.4
theta<-4
T<-30
h<-0.01
t<-seq(0,T,h)
n<-length(t)
M<-1000
est<-rep(0,M)

 for( j in 1:M)
 {
   W<-c(0,cumsum(rnorm(n-1,0,1)*sqrt(h)))
   y<-rep(y0,n)


for(i in 1:(n-1))
{
  y[i+1]<-y[i]+a*t[i]*h-b*y[i]*h+theta*(W[i+1]-W[i])
}

delta<-1/n
X<-y
M<-(a/b)+(X-(a/b))*exp(-b*delta)
est[j]<-sqrt((2*b)/(1-exp(-2*b*delta))*(1/n)*sum(X[2:n]-M[1:(n-1)]))
 }

hist(sqrt(n)*(est-theta),nclass = 40,freq = FALSE)

Please note that I do not need to get the line graph of the histogram itself at all. I just need to add other normal distribution to compare it with the histogram. This new normal distribution is defined by:

 x<-seq(-1,1,length=1000)
    y<-dnorm(x,mean = 0,sd=sqrt(theta^2/2))
    lines(x,y)

But I did not get any curve plot added to the graph of the density function defined above.

just I can't get the density distribution

I did understand what's the problem? Thank you for you help!

edited Apr 08 '18 at 16:34

pjs

18,696
4
27
56

asked Apr 07 '18 at 23:34

user8810618

1

like this? `curve(dnorm(x, mean = mean(sqrt(n)*(est-theta)), sd=sqrt(theta^2/2)), add = TRUE)` – rawr Apr 07 '18 at 23:42
@rawr, thank you very much !! – user8810618 Apr 08 '18 at 00:47
@rawr, no the mean is not as you write. the mean is null as I posted in my code. Here I didn't want to plot the normal density of the distribution it self no.But I need to compare the histogram distribution to other one difined by `dnorm(x,mean = 0,sd=sqrt(theta^2/2))` – user8810618 Apr 08 '18 at 14:59
@Maurits Evers, please note that's not duplicated! look to the edited one. – user8810618 Apr 08 '18 at 15:07
well then you can't fit it on your figure because of the x range – rawr Apr 08 '18 at 15:20
@rawr, how should I modify the x range so? – user8810618 Apr 08 '18 at 15:38
either center your histogram at 0 or use the calculated mean for the density like I did – rawr Apr 08 '18 at 15:50
@rawr, please help me by writing an R code. – user8810618 Apr 08 '18 at 16:19
@rawr, I tried using `axis(side=1,at=foo$mids)` but no way! – user8810618 Apr 08 '18 at 20:59
`hist(scale(sqrt(n)*(est-theta), scale = FALSE),nclass = 40,freq = FALSE); curve(dnorm(x, mean = 0, sd=sqrt(theta^2/2)), add = TRUE)` – rawr Apr 08 '18 at 22:40
@rawr, I did this many times choosing x as `x<-seq(-1,1,length=100)` and it doesn't work at all – user8810618 Apr 08 '18 at 22:46

Can't get a different density function to the histogram distribution plotted

0 Answers0