Is the normalpdf function providing correct values?

Helloworld

I don't know if I am missing something here but the values of the normalpdf function don't look correct. Can someone please help?
NormalPdf(1,0,1)=0.24197
Shouldn't the correct value be 0.84134

Desa

That value is correct. Why are you expecting 0.84134?
http://www.mathstudio.net/manual/NormalPDF

APB

I think you're confusing it with the CDF. This would be 0.84134 if you were integrating from minus large number to 1 with unit variance centred at zero, i.e.

NormalCdf(-100,1,0,1)

I notice that

NormalCdf(-100,1)

doesn't currently work, although the manual page suggests that it should with default [0,1] for the last two parameters.

Edit: Indeed I meant minus infinity. I should have put e.g. not i.e. above as there is nothing special about -100. As Stephen confirms below there's negligible density 10 standard errors away from the mean. A value of 100 times the standard error was overkill and in fact 5 would have done the job.

Desa

Thank you for pointing that out. That is fixed for the next update now.

Helloworld

Thanks. I was using the wrong function.

Helloworld

I check out the manual but didn't find an explanation for -100. Experimenting -10 does the same result! So I guess -100 is acting as -oo as explained above. I am very grateful for the help. I believe some examples are necessary in the help file. Thanks again.

Shakey

The reason why -10 works as well as -100 is because the function decays very fast, on the order of exp(-x^2/2), so by the time you get to -10 you are looking at exp(-50) which is approximately 10^(-22).

Stephen

Howdy, Stranger!

Categories

In this Discussion