pyuncertainnumber.calibration.pdfs¶

@author: Mukesh K. Ramancha

A collection of common probability distributions

Classes¶

`ProbabilityDensityFun`	Blueprint for other classes.
`Uniform`	Uniform continuous distribution
`HalfNormal`	Half Normal distribution with zero mean
`Normal`	Normal distribution
`TruncatedNormal`	Truncated Normal distribution
`MultivariateNormal`	Multivariate Normal distribution

Module Contents¶

class pyuncertainnumber.calibration.pdfs.ProbabilityDensityFun¶

Bases: abc.ABC

Blueprint for other classes. Base class. Abstract class is not a concrete class, it cannot be instantiated

abstractmethod generate_rns(N)¶

Method to generate ‘N’ random numbers

Parameters:: N (int) – number of random numbers needed.
Return type:: numpy array of size N

abstractmethod log_pdf_eval(x)¶

Method to compute log of the pdf at x

Parameters:: x (float) – value where to evalute the pdf.
Return type:: float - log of pdf evaluated at x.

class pyuncertainnumber.calibration.pdfs.Uniform(lower=0, upper=1)¶

Bases: ProbabilityDensityFun

Uniform continuous distribution

lower = 0¶

upper = 1¶

generate_rns(N)¶

Method to generate ‘N’ random numbers

Parameters:: N (int) – number of random numbers needed.
Return type:: numpy array of size N

log_pdf_eval(x)¶

Method to compute log of the pdf at x

Parameters:: x (float) – value where to evalute the pdf.
Return type:: float - log of pdf evaluated at x.

class pyuncertainnumber.calibration.pdfs.HalfNormal(sig=1)¶

Bases: ProbabilityDensityFun

Half Normal distribution with zero mean

sig = 1¶

generate_rns(N)¶

Method to generate ‘N’ random numbers

Parameters:: N (int) – number of random numbers needed.
Return type:: numpy array of size N

log_pdf_eval(x)¶

Method to compute log of the pdf at x

Parameters:: x (float) – value where to evalute the pdf.
Return type:: float - log of pdf evaluated at x.

class pyuncertainnumber.calibration.pdfs.Normal(mu=0, sig=1)¶

Bases: ProbabilityDensityFun

Normal distribution

mu = 0¶

sig = 1¶

generate_rns(N)¶

Method to generate ‘N’ random numbers

Parameters:: N (int) – number of random numbers needed.
Return type:: numpy array of size N

log_pdf_eval(x)¶

Method to compute log of the pdf at x

Parameters:: x (float) – value where to evalute the pdf.
Return type:: float - log of pdf evaluated at x.

class pyuncertainnumber.calibration.pdfs.TruncatedNormal(mu=0, sig=1, low=-np.Inf, up=np.Inf)¶

Bases: ProbabilityDensityFun

Truncated Normal distribution

mu = 0¶

sig = 1¶

low¶

up¶

generate_rns(N)¶

Method to generate ‘N’ random numbers

Parameters:: N (int) – number of random numbers needed.
Return type:: numpy array of size N

log_pdf_eval(x)¶

Method to compute log of the pdf at x

Parameters:: x (float) – value where to evalute the pdf.
Return type:: float - log of pdf evaluated at x.

class pyuncertainnumber.calibration.pdfs.MultivariateNormal(mu=np.zeros(2), E=np.identity(2))¶

Bases: ProbabilityDensityFun

Multivariate Normal distribution

mu¶

E¶

d¶

logdetE¶

Einv¶

generate_rns(N)¶

Method to generate ‘N’ random numbers

Parameters:: N (int) – number of random numbers needed.
Return type:: numpy array of size N

log_pdf_eval(x)¶

Method to compute log of the pdf at x

Parameters:: x (float) – value where to evalute the pdf.
Return type:: float - log of pdf evaluated at x.