CODEWITHSUNDEEP
pythonscipy
BhishanPoudel
BhishanPoudel

Reputation: 17164

How to create lognormal numbers between a and b in scipy1.0?

I am using a computer with scipy version 1.0 and can not update the module. I need to create N lognormal points between a and b, How to create the random numbers?

For scipy 1.4

from scipy import stats
x = stats.loguniform.rvs(1,10,size=100,random_state=100)

How to do in scipy1.0

Scipy 1.0 does not have stats.lognormal it only has stats.lognorm with shape parameter.

Any methods using numpy or random modules with also be welcome it need not be only using scipy.

Upvotes: 0

Views: 180

Answers (1)

Warren Weckesser
Warren Weckesser

Reputation: 114891

The distribution scipy.stats.loguniform is just the new name of scipy.stats.reciprocal, and reciprocal is available in SciPy 1.0. So if loguniform is what you need, you can use reciprocal instead.

Note, however, that the log-uniform distribution is not the same as the log-normal distribution, so it is a bit confusing that the question asked for lognormal random variates, but suggests loguniform is an acceptable distribution. It sounds like what you really want is a truncated log-normal distribution.

Upvotes: 1

Related Questions