Monte Carlo

[jimmy1899]

Hello,

I have had a look at the Monte Carlo examples. I think I will store the simulation results into a matrix.

However I would have liked to learn how to program some kind of recursive algorithm with EViews.

What's a good way to program such a model : Y(t) = a*Y(t-1)+b+nrnd ?

Thanks for you help !
Jimmy

[EViews Gareth]

You'll need to seed Y_1 at some value, after that it is relatively simple:

Code: Select all

'seed Y_1
smpl @first @first
series y = nrnd

'calculate the rest of the values
smpl @first+1 @last
y = a*y(-1) + b + nrnd


[jimmy1899]

Thank you very much Gareth

"smpl @first @first" and "smpl @first+1 @last" are delimiters

To set the first element to 5, I could also have written

Code: Select all

smpl @first @first
series y = 5

This looks pretty easy indeed

Thanks !

[kzmh78]

Hello,

When monte carlo program generate a sequence of random numbers,
almost of examples had used a Nornal disitrubution,mean is 0, and variance is 1(using command @rnorm or nrnd).

However, I want to use a nornal disitrubution,mean is 2, and variance is 3.

Does EViews have any commands?

[startz]

Code: Select all

2+sqr(3)*nrnd

[mvilla29]

Hi
I'm triyin to generate a random number serie, distribited logistic with mean 85.22 and Standard Deviation 18.44, with 260 observations but I don´t know how to do this. I want to to do a Montecarlo experiment with 10.000 iteration like this but I don't know how to do this. Can you help me, please?

Thks!

[EViews Gareth]

The @rlogistic function creates random variables from the logistic distribution.

[mvilla29]

The @rlogistic function creates random variables from the logistic distribution.

Ok, but, how can i asign a mean of 85.22 and a standard deviation of 18.4?

[EViews Gareth]

To generate a series with mean X and scale parameter Y, you would do:

Code: Select all

series z = x + y*@rlogistic


[mvilla29]

Thank you Gareth!!

[startz]

To generate a series with mean X and scale parameter Y, you would do:

Code: Select all

series z = x + y*@rlogistic


Actually, not quite. The standard deviation of the logistic isn't 1.0. So the scale doesn't give you the standard deviation.

[EViews Gareth]

I didn't say anything about std dev.

[EViews Gareth]

A quick bit of Google-fu indicates that to get a mean of m and standard deviation of s, you'd do:

Code: Select all

series y = m + s*@sqrt(3/(16*@atan(1)^2))*@rlogistic