How to develope simulation for gev using Log Likelihood estimate

[mnuramalina]

I have developed a simulation for GEV distribution by using log likelihood estimate. However, there are some errors on that coding.

From that code, I expect every parameter (location, scale, and shape) should be come out in matrices not an integer.

Hope someone can help me.Thank you.

[Fred_Kohlhepp]

I kind of got lost in what you're trying to do, but it's clear that you need to pay better attention to arrays and subscripts:

• Mathcad default subscript starts at ORIGIN = 0, you can change that--since I and j both started at 1, you need to redefine ORIGIN or all your vectors will start with a 0.
• You can define a whole column of an array, the keystroke sequence is "cntrl ^"
• For double subscripts the first is row, the second column

I started making changes.  See if that start inspires you.

[mnuramalina]

Thank you for the notes of arrays and subscripts info.

Actually, I am doing for simulation by using Monte Carlo simulation for GEV distribution by MLE method. This method of simulation uses repeated random sampling to generate simulated data to use with a mathematical model.

I managed to estimate the parameter of GEV distribution by MLE method from the specific data. but, I had problem on estimate parameter by simulated data.

I have tried the idea that you stated. It was good info. But, I still on that problem.

**I am not sure why I can not attach my programming in here.

[Werner_E]

> **I am not sure why I can not attach my programming in here.

This forum system makes it unnecessarily difficult. You have to chose "Use advanced editor" at the upper right. Then you are offered (at the lower right) the option to attach files.

[LucMeekes]

The other option to attach your file is, after you've created a response, click the 'actions' options at the lower left side and choose 'edit' from the list. Then you will find an 'attach' option at the lower right side.

[mnuramalina]

Thank you for all helpers.

Here are my attachments for my previous comment.

[-MFra-]

Below is my attempt to fix your problem. There are imaginary or complex values. The final function is actually a vector, since its values also depend on the index j. How to maximize?

[mnuramalina]

Thank you for the idea, F.M.. in Mathcad we used 'Maximize' to maximize it and there is an error on that.

I am not sure if there have other way to maximize this loglikelihood function.