tech.nmfin.portfoliooptimization.lai2010

Class Lai2010NPEBModel

java.lang.Object

tech.nmfin.portfoliooptimization.lai2010.Lai2010NPEBModel

public class Lai2010NPEBModel
extends Object

The Non-Parametric Empirical Bayes (NPEB) model described in the reference computes the optimal weights for asset allocation.

See Also:

• "Tze Leung Lai, Haipeng Xing, Zehao Chen. "Mean-variance portfolio optimization when means and covariances are unknown," Annals of Applied Statistics 2011, Vol. 5, No. 2A, p. 798-823"
• MATLAB code

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	Lai2010NPEBModel.OptimalWeights

Constructor Summary

Constructors

Constructor and Description
Lai2010NPEBModel(MVOptimizer optimizer, int nBootstraps) Constructs an instance of the model using MomentsEstimatorLedoitWolf as the default estimator of covariance matrix, and assuming IID for each individual asset for re-sampling.
Lai2010NPEBModel(MVOptimizer mvOptimizer, int nBootstraps, ReturnsMoments.Estimator returnsMomentsEstimator, ReturnsResamplerFactory resamplerFactory, CetaMaximizer cetaMaximizer) Constructs an instance of the model.

Method Summary

Modifier and Type	Method and Description
Lai2010NPEBModel.OptimalWeights	optimalWeights(Matrix returns, double lambda) Computes the weights based on given historical returns and the risk-aversion index λ.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Lai2010NPEBModel

public Lai2010NPEBModel(MVOptimizer optimizer,
                        int nBootstraps)

Constructs an instance of the model using MomentsEstimatorLedoitWolf as the default estimator of covariance matrix, and assuming IID for each individual asset for re-sampling.

Parameters:

optimizer - an optimizer to compute optimal weights

nBootstraps - the number of bootstraps to be used

Lai2010NPEBModel

public Lai2010NPEBModel(MVOptimizer mvOptimizer,
                        int nBootstraps,
                        ReturnsMoments.Estimator returnsMomentsEstimator,
                        ReturnsResamplerFactory resamplerFactory,
                        CetaMaximizer cetaMaximizer)

Constructs an instance of the model.

Parameters:

mvOptimizer - an optimizer to compute optimal weights

nBootstraps - the number of bootstraps to be used

returnsMomentsEstimator - an estimator of covariance matrix

resamplerFactory - the resampling mechanism

cetaMaximizer - a maximizer for the C-eta function

Method Detail

optimalWeights

public Lai2010NPEBModel.OptimalWeights optimalWeights(Matrix returns,
                                                      double lambda)

Computes the weights based on given historical returns and the risk-aversion index λ.

Parameters:

returns - the historical returns matrix

lambda - the risk-aversion index

Returns:

the optimal weight vector