FeatureSelection: Feature Selection

In Bjerring/BEWESO: BEWESO

Description

Use feature selection to cluster assets into correlated groups following Bjerring et al. 2016. This is a variation, where assets are choses according to their Omega ratio (See Keating and Shadwick, 2002) and Rf is defined as the discounted annual holding fees.

Usage

  FeatureSelection(returns,Cno,fees,data.length,OKlist)

Arguments

returns	matrix holding returns of n assets
Cno	Number of groups in the hierarchical clustering
fees	annual expenses for holding an asset
data.length	number of data points that the correlation matrix should be estimated over
OKlist	vector of asset names which have to be in the reduces asset universe

Examples

data(ETFdata)
data(ETFdescription)

fees <- ETFdescription[,5]
FeatureSelection(returns=ETFdata,Cno=5,fees)

Bjerring/BEWESO documentation built on May 6, 2019, 7:56 a.m.