fms2: Covariance Estimation by modified PCA

In monkeypicked/aace: Covariance estimation

Description

Covariance Estimation by modified PCA

Usage

fms2(x, weight = seq(1, 3, length = nobs), center = TRUE, frac.var = 0.5,
  iter.max = 1, nfac.miss = 1, full.min = 20, reg.min = 40,
  sd.min = 20, quan.sd = 0.9, tol = 0.001, zero.load = FALSE,
  range.factors = c(20, 20), lambda = 0, minunique = 0.02,
  shrinkb = 0.3, shrinkv = shrinkb, shrinkr = 0.9, ...)

Arguments

x	matrix or dataframe of timeseries returns
weight	weights in estimation
center	flag to center
frac.var	controls auto-selection of number of factord
iter.max	maximum number of iterations
nfac.miss	number of factors to estimate if data is missing
full.min	minimum acceptable number of NA-free columns
reg.min	minimum dates to do regression
sd.min	minimum dates to estimate vol
quan.sd	missing vol assigned this quantile
tol	estimation tolerance
zero.load	flag to use zero loadings for columns with missing
range.factors	range of factors to estimate, as a function of valid data length
lambda	exponent on eigenvalue for shrinkage
minunique	minimum uniqueness
shrinkb	shrinkage for factor 1
shrinkv	shrinkage for vol
shrinkr	shrinkage for regressed loadings

Value

list(loadings fmp hpl method full uniqueness sdev qua weight call)

Details

more detail on the underlying algorithm may be found in documentation for BurStFin

Author(s)

Giles Heywood from Pat Burns original

monkeypicked/aace documentation built on May 23, 2019, 6:10 a.m.