idio.seg: Multiple change-point detection for the idiosyncratic...

In factorcpt: Simultaneous Change-Point and Factor Analysis

Description

First generating the panel of statistics via wavelet-based filtering of the estimated idiosyncratic components, it applies the Double CUSUM Binary Segmentation in combination with the bootstrap generated thresholds to estimate the multiple change-points in the idiosyncratic components.

Usage

idio.seg(gfm, q, scales = NULL, diag = TRUE, sig.lev = 0.05, rule = NULL, B = 200,
        p = NULL, dw = NULL, mby = NULL, tby = NULL, do.parallel = TRUE)

Arguments

gfm	a get.factor.model object with estimates of the factor structure
q	the number of factors
scales	see scales in factor.seg.alg
diag	see idio.diag in factor.seg.alg
sig.lev	see sig.lev in factor.seg.alg
rule	the height of a binary tree for change-point analysis, see the Appendix of Barigozzi, Cho & Fryzlewicz (2016)
B	the size of bootstrap samples
p	see p in factor.seg.alg
dw	see dw in factor.seg.alg
mby	see dmby in func_dc_by
tby	see dtby in func_dc_by
do.parallel	see do.parallel in factor.seg.alg

Value

tree	a list containing the binary tree grown for change-point analysis on the idiosyncratic components; each node contains its index, the start and end of the segment as well as the estimated change-point, the test statistic and the proportion of bootstrap statistics smaller than the test statistic.
est.cps	a vector of change-points estimated for the idiosyncratic components; adjusted for the bias introduced by the wavelet filtering
p	the reciprocal of the average block size selected for the estimated idiosyncratic components

Author(s)

Haeran Cho

References

M. Barigozzi, H. Cho and P. Fryzlewicz (2016) Simultaneous multiple change-point and factor analysis for high-dimensional time series, Preprint.

H. Cho (2016) Change-point detection in panel data via double CUSUM statistic. Electronic Journal of Statistics. 10: 2000-2038.

factorcpt documentation built on May 2, 2019, 8:15 a.m.