Parameter influence estimation using monte carlo

Question

Nur Farahin Che Nordin 2024-8-14

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2145359-parameter-influence-estimation-using-monte-carlo

回答： Rahul 2024-9-2

Hi,

I have 14 parameters of groundwater, Ca, Mg, pH, Cl, NO3, TD, EC.. etc. I want to estimate the influence of each parameters for groundwater quality using monte carlo but had no idea about the coding. I try to look out for some coding but really didnt get it.

Please help

2 个评论
显示无隐藏无

John D'Errico 2024-8-14

How is this a question about MATLAB, except that you would like someone to teach you both about a method you vaguely describe, and about the code to perform that operation?

I would strongly suggest you first figure out exactly what and how you need to do this. Don't worry about MATLAB, about code. Then and only when you know what you need to do, should you worry about code to do it.

Sam Chak 2024-8-14

Hi @Nur Farahin Che Nordin,

Please complete the following two tasks:

Identify the mathematical expression that describes the relationship between groundwater quality and the 14 parameters. If the expression or equation can be expressed in closed form, you can create graphs over a defined operating range using the 'plot()' command.
Define your Monte Carlo-style experiments, as they rely on repeated random sampling to obtain numerical results. How do you intend to perform random sampling? Will you impose mathematical constraints on the randomness of the sampling?

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Rahul 2024-9-2

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2145359-parameter-influence-estimation-using-monte-carlo#answer_1508939

在 MATLAB Online 中打开

Hi @Nur Farahin Che Nordin,

I understand that you have 14 parameters based on which you want to run the 'Monte Carlo' Simulation to estimate the influence of these parameters on the groundwater.

'Monte Carlo' estimation is based on generating random samples from your data, running your model, and analyzing the results.

You can refer to the below provided code, where I have created a simple 'Monte Carlo' linear model based on ‘mean’ estimation:️

numSimulations = 1000; % You can adjust the number of simulations accordingly
% Assuming 'data' to be the variable of Matrix containing data from all 14 parameters
% as columns.
% Monte Carlo estim
for i = 1:numSimulations
    % Randomly sample indices
    indices = randi(size(data, 1), size(data, 1), 1);
    sampleData = data(indices, :);
    
    % Simple model: Mean calculation
    qualityIndicator = mean(sampleData, 1);
    results(i, :) = qualityIndicator;
end
% Analyze and Display results accordingly
meanResults = mean(results);
stdResults = std(results);
disp('Mean influence of each parameter:');
disp(meanResults);
disp('Standard deviation of influence:');
disp(stdResults);
figure;
bar(meanResults);
hold on;
errorbar(1:length(meanResults), meanResults, stdResults, '.');
xlabel('Parameter');
ylabel('Influence');
title('Influence of Parameters on Groundwater Quality');