starvine.bvcopula.pc_base.PairCopula Class Reference

Stores bivariate data for pair copula construction. More...

Inheritance diagram for starvine.bvcopula.pc_base.PairCopula:

Public Member Functions
def	__init__ (self, x, y, weights=None, kwargs)
	Bivariate data set init. More...
def	rank (self, method=0)
	rank transfom the data More...
def	rankInv (self)
	Inverse rank transform data back to original scale.
def	setTrialCopula (self, family)
def	empKTau (self)
	Returns emperical kendall's tau of rank transformed data. More...
def	empSRho (self)
	Returns emperical spearman rho, the rank correlation coefficient. More...
def	empPRho (self)
	Returns linear correlation coefficient, pearson's rho. More...
def	copulaTournament (self, criterion='AIC', kwargs)
	Determines the copula that best fits the rank transformed data based on the AIC criterion. More...
def	fitCopula (self, copula, thetaGuess=(None, None,))
	fit specified copula to data. More...
def	rotateData (self, u, v, rotation=-1)
	Rotates the ranked data on the unit square. More...
def	setRotation (self, rotation=0)
	Set the copula's orientation: 0 == 0 deg 1 == 90 deg rotation 2 == 180 deg rotation 3 == 270 deg rotation Allows for modeling negative dependence with the frank, gumbel, and clayton copulas (Archimedean Copula family is non-symmetric)

Public Attributes
	copulaParams
	id
	x
	y
	v
	weights
	rankMethod
	u
	trialFamily
	copulaBank
	pval_
	copulaModel
	UU
	VV
	rotation

Detailed Description

Stores bivariate data for pair copula construction.

Contains methods to:

• rank data transform
• rotate data
• remove nan or inf data points
• plot bivarate data
• fit copula to bivariate (ranked) data
• compute basic bivariate statistics (eg. kendall's tau)

Note: Depends on pandas for some useful statistical and plotting functionality.

Constructor & Destructor Documentation

__init__()

def starvine.bvcopula.pc_base.PairCopula.__init__	(		self,
			x,
			y,
			weights = None,
			kwargs
	)

Bivariate data set init.

Parameters

x	np_1darray first marginal data set
y	np_1darray second marginal data set
weights	np_1darray (optional) data weights normalized or unormalized weights accepted Note: len(u) == len(v) == len(weights)

Member Function Documentation

copulaTournament()

def starvine.bvcopula.pc_base.PairCopula.copulaTournament	(		self,
			criterion = 'AIC',
			kwargs
	)

Determines the copula that best fits the rank transformed data based on the AIC criterion.

All Copula in self.trialFamily set are considered.

empKTau()

def starvine.bvcopula.pc_base.PairCopula.empKTau

(

self

)

Returns emperical kendall's tau of rank transformed data.

Returns

float Kendall's tau rank correlation coeff

empPRho()

def starvine.bvcopula.pc_base.PairCopula.empPRho

(

self

)

Returns linear correlation coefficient, pearson's rho.

Returns

float pearson's correlation coefficient

empSRho()

def starvine.bvcopula.pc_base.PairCopula.empSRho

(

self

)

Returns emperical spearman rho, the rank correlation coefficient.

Returns

float Spearman's rank correlation coeff

fitCopula()

def starvine.bvcopula.pc_base.PairCopula.fitCopula	(		self,
			copula,
			thetaGuess = (None, None, )
	)

fit specified copula to data.

Parameters

copula	CopulaBase Copula instance
thetaGuess	tuple (optional) initial guess for copula params

Returns

(copula type string, fitted copula params np_array)

rank()

def starvine.bvcopula.pc_base.PairCopula.rank	(		self,
			method = 0
	)

rank transfom the data

Parameters

method

int if == 0: use standard rank transform, else: use CDF data transform.

rotateData()

def starvine.bvcopula.pc_base.PairCopula.rotateData	(		self,
			u,
			v,
			rotation = -1
	)

Rotates the ranked data on the unit square.

Parameters

u	Ranked data vector
v	Ranked data vector
rotation	int 1==90deg, 2==180deg, 3==270, 0==0deg

---

The documentation for this class was generated from the following file:

• starvine/bvcopula/pc_base.py