Skip to content

How to plot PDF which is in the form of orthogonal polynomial

2 messages · Yong Wang, Spencer Graves

Original
Yong Wang
# 
Dear all
using the orthogonal polymial on a set of data, I get an approximate 
density
which basically is in the form: exp(-polynomial),
  as you know, the parameters are the converged coeeficients.
obviously, It is hard, if not impossible, to use the inverse CDF method to 
get
a sample and then plot density. then how can I plot the approximated 
density in
order to have a graphical comparision with the real data's histogram.

any hint is appreciated

thanks
Spencer Graves
# 
Have you considered "cumsum(exp(-polynomial))"? 

      Consider for example the following: 

library(polynom)
b. <- polynomial(1:5)

x <- seq(-2, 2, length=101)
b.p <- predict(b., x)

plot(x, b.p)
plot(x, exp(-b.p))

cdf <- cumsum(exp(-b.p))
plot(cdf/cdf[101])
  
      hope this helps.  spencer graves
Yong Wang wrote: