Monte Carlo Simulation/Markov Chains

9 views (last 30 days)
Eric on 12 May 2014
Answered: Ragini Gupta on 22 Nov 2017
I was assigned a computer project regarding Markov Chains and simulations. Trouble is, I've never taken a probability course in my life and am having some heavy difficulty starting the project and was wondering if anybody would be able to assist me. I have a forward and reverse transition rate for a simple two state model but am not sure what to do with these equations.

Answers (1)

Ragini Gupta
Ragini Gupta on 22 Nov 2017
Hi all,
Since I am new to MCMC simulation I am facing a similar problem.I have to simulate a smart meter data for a week's time using Markov chain model. Now,I need to run the markov model to generate a new Smart meter value for each day (i.e. 7 readings). However, from the following code I know how to train the model untill using the dhmm_em function.But I do not understand how to use this trained model and generate a new value for the next state.
Here is the code I have so far:
filename = 'C:\Users\profile\Desktop\SmartMeter.xlsx'; Furnace=xlsread(filename,'E:E'); %H1D1 x=length(Furnace); hist(Furnace) edges = linspace(min(Furnace),max(Furnace),10); % generates a row vector of 10 linearly equally spaced points between min % and max furnace values
% edges are the end points of each bin
[counts,bins] = histc(Furnace, edges);
figure bar(edges,counts,'histc') O = 10; % output symbols Q = 5; % number of states
prior0 = bins; T=10; nex=20; t1 = datetime(2017,1,1,24,0,0); t2 = datetime(2017,1,7,24,0,0); t = t1:hours(24):t2; s=length(t);
%Generate Sequence of Dates and Time with a step size of 15 minutes fileID = fopen('C:\Users\profile\Desktop\raginiData5.txt','w'); for t = t1:hours(24):t2 transmat0 = mk_stochastic(rand(Q,Q)); obsmat0 = mk_stochastic(rand(Q,O)); %Now we sample nex=20 sequences of length T=10 each from this model, to use as training data. data = dhmm_sample(prior0, transmat0, obsmat0, nex, T); %SAMPLE_DHMM Generate random sequences from a Hidden Markov Model with discrete outputs. %Each row of data is an observation sequence of length T. % initial guess of parameters prior1 =normalise(rand(Q,1)); transmat1 = mk_stochastic(rand(Q,Q)); obsmat1 = mk_stochastic(rand(Q,O));
% improve guess of parameters using EM algorithm (using the
% maximum likelihood of parameters in statistical model
[LL, prior2, transmat2, obsmat2] = dhmm_em(data, prior1, transmat1, obsmat1);
DateString = datestr(t);
fprintf(fileID,'%s %f \n',DateString);
end fclose(fileID);
Please can anyone advise me on the next step.
Thanks in advance.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!