What are the possible ways to generate a random time series?

Hi everyone,

I require generating random times to test the computation. However, I got totally different results when I repeated my experiment for 1000 and, 10000 random time values. Here is the script that I used to generate random times.

NUMBER_RANDOM_TIMES = 10000;
SECONDS_PER_DAY = 24*60*60;
START_DATE = '2015-07-01';
END_DATE   = '2021-12-31';
% Algorithm
startDateNum = datenum(START_DATE,'yyyy-mm-dd');
endDateNum   = datenum(END_DATE,  'yyyy-mm-dd');
dayRange = endDateNum - startDateNum;
secondsRange = SECONDS_PER_DAY*dayRange;
randomNumberOfSeconds = randi(secondsRange,NUMBER_RANDOM_TIMES,1);
randomDatenums = startDateNum + randomNumberOfSeconds/SECONDS_PER_DAY;
AA = datevec(randomDatenums);

I think maybe the random times generated by this method are biased. May someone share their thoughts and suggest other possible ways to generate random time's bounded between the start and end dates.

Thank you!

[Chunru]

There is no problem in the above random time generate. It is expected that the results (not sure what results are referred to) will be different with 1000 trials and 10000 trials. In theory, the results (some statistics ???) will be converging when number of trials are bit enough.

Answer 1

[Muskan]

Hi Andi,

As per my understanding of the question when generating random numbers, the results can vary each time you run the experiment due to the nature of the random number generation algorithm.

The difference in results between 1000 and 10000 random time values could be due to the larger sample size in the latter case.

You can further refer to the following answer and documentation for a better understanding:

how to generate random time format data in MATLAB? - MATLAB Answers - MATLAB Central (mathworks.com)

https://in.mathworks.com/help/matlab/ref/timeseries.html

I hope the above information helps resolve your query.

Thanks