How to draw samples from lmfit model?

Question

I have a skewed Gaussian model in lmfit that fits my data. Now I would like to draw a sample from it, but I could nowhere in the documentation find how? Is the correct approach to simply implement the model function yourself in my case the skewed normal distribution or is there a function in lmfit for this?

My code:

model = SkewedGaussianModel()

params = model.make_params(amplitude=60, center=30, sigma=10, gamma=0)

result = model.fit(y, params, x=x)
print(result.fit_report())
plt.plot(x, result.best_fit)
plt.show()
# something like this
print(result.model.eval(random.random())

score 1 · Answer 1 · edited Dec 12 '17 at 11:21

Until somebody can find the function, or confirm the nonexistence of it here is how I did it:

def pdf(x):
    return 1/sqrt(2*pi) * exp(-x**2/2)

def cdf(x):
    return (1 + erf(x/sqrt(2))) / 2

def skew(x,e=0,w=1,a=0):
    t = (x-e) / w
    return 2 / w * pdf(t) * cdf(a*t)
    # You can of course use the scipy.stats.norm versions
    # return 2 * norm.pdf(t) * norm.cdf(a*t)

copied from this answer

score 1 · Accepted Answer · answered Jul 04 '17 at 17:07

The definition for skewed Gaussian used in lmfit is given at http://lmfit.github.io/lmfit-py/builtin_models.html#skewedgaussianmodel with code at https://github.com/lmfit/lmfit-py/blob/master/lmfit/lineshapes.py#L213.

I believe that what you are looking for is the "inverse transform sampling". See http://www.nehalemlabs.net/prototype/blog/2013/12/16/how-to-do-inverse-transformation-sampling-in-scipy-and-numpy/ for hints on how to do that. There is not a built-in method for this in lmfit, as lmfit does not necessarily assert that the model being fitted is a probability distribution function. It might be worth considering adding such a capability.

Thank you for these resources. This is the right answer. – Hakaishin Jul 22 '17 at 18:55 — Hakaishin, Jul 22 '17 at 18:55

How to draw samples from lmfit model?

2 Answers2