wdm Namespace Reference

Weighted dependence measures. More...

Classes
class	Indep_test

Functions
double	wdm (std::vector< double > x, std::vector< double > y, std::string method, std::vector< double > weights=std::vector< double >(), bool remove_missing=true, std::vector< int > seeds=std::vector< int >())
double	wdm (const Eigen::VectorXd &x, const Eigen::VectorXd &y, std::string method, Eigen::VectorXd weights=Eigen::VectorXd(), bool remove_missing=true, std::vector< int > seeds=std::vector< int >())
Eigen::MatrixXd	wdm (const Eigen::MatrixXd &x, std::string method, Eigen::VectorXd weights=Eigen::VectorXd(), bool remove_missing=true, std::vector< int > seeds=std::vector< int >())

Detailed Description

Weighted dependence measures.

Function Documentation

wdm() [1/3]

Eigen::MatrixXd wdm::wdm ( const Eigen::MatrixXd &  x, std::string  method, Eigen::VectorXd  weights = Eigen::VectorXd(), bool  remove_missing = true, std::vector< int >  seeds = std::vector<int>()  )

inline

Calculates all pairwise dependence measures between matrix columns.

Parameters

x	matrix with observations in rows and variables in columns.
method	any method name accepted by the vector overload.
weights	optional case weights for the rows.
remove_missing	whether each pair should remove rows containing a NaN.
seeds	optional seeds for random Chatterjee predictor-tie breaking.

Returns

A matrix of pairwise dependence measures. Chatterjee matrices are generally asymmetric; all other supported measures produce symmetric matrices.

Exceptions

std::runtime_error

if x has exactly one column or a pairwise call rejects its inputs.

wdm() [2/3]

double wdm::wdm ( const Eigen::VectorXd &  x, const Eigen::VectorXd &  y, std::string  method, Eigen::VectorXd  weights = Eigen::VectorXd(), bool  remove_missing = true, std::vector< int >  seeds = std::vector<int>()  )

inline

Calculates a dependence measure for two Eigen vectors.

Parameters

x,y	input vectors of equal length.
method	any method name accepted by the std::vector overload.
weights	optional case weights subject to the same validation as the std::vector overload.
remove_missing	whether to remove rows containing a NaN.
seeds	optional seeds for random Chatterjee predictor-tie breaking.

Returns

The result of the corresponding std::vector overload.

Exceptions

std::runtime_error

under the same conditions as that overload.

wdm() [3/3]

double wdm::wdm ( std::vector< double >  x, std::vector< double >  y, std::string  method, std::vector< double >  weights = std::vector<double>(), bool  remove_missing = true, std::vector< int >  seeds = std::vector<int>()  )

inline

Calculates a weighted or unweighted dependence measure.

Parameters

x,y	input vectors of equal length. For Chatterjee's xi, x is the predictor and y is the response.
method	the dependence measure; see details for possible values.
weights	optional case weights. Nonempty weights must match the input length, be finite and nonnegative, and have a positive sum. Their scale does not affect the result, and zero-weight rows are ignored.
remove_missing	if true, rows containing a NaN are removed; otherwise a NaN raises an exception.
seeds	optional seeds for random Chatterjee predictor-tie breaking.

Available methods:

• "pearson", "prho", "cor": Pearson correlation
• "spearman", "srho", "rho": Spearman's \( \rho \)
• "kendall", "ktau", "tau": Kendall's \( \tau \)
• "blomqvist", "bbeta", "beta": Blomqvist's \( \beta \)
• "hoeffding", "hoeffd", "d": Hoeffding's \( D \)
• "chatterjee", "cxi", "xi": Chatterjee's \( \xi \)

Note

Chatterjee's xi is asymmetric. Predictor ties are broken without consulting the response; use seeds for a reproducible ordering.

Returns

The requested estimate, or NaN if missing-value removal leaves fewer than two observations (five for Hoeffding's D).

Exceptions

std::runtime_error

for size mismatches, invalid weights, unknown methods, disallowed missing or insufficient input, or an undefined Chatterjee estimate with a constant response.