Aug-06-2019, 10:41 PM
As far as I understood, you want to drop the original sample and generate a new one based on
obtained values of mean, std, skew. These are only three parameters, so it is better
to know (at least) class of allowed distributions; For example, normal distribution is fully determined
by mean and std values. So, if you knew that your data came from normal distribution, it
would be sufficient to just sample from the normal distribution with specific parameters,
e.g. using
Take a look at MCMC approach, it is usually used when needed to approximate a distribution.
obtained values of mean, std, skew. These are only three parameters, so it is better
to know (at least) class of allowed distributions; For example, normal distribution is fully determined
by mean and std values. So, if you knew that your data came from normal distribution, it
would be sufficient to just sample from the normal distribution with specific parameters,
e.g. using
numpy.random.randn.Take a look at MCMC approach, it is usually used when needed to approximate a distribution.