dynamicsconv: dynamicsconv

In vancebee/MarkovSCD:

Description

Calcualte delta matrix for increasingly larger subsets of the data and compere to the delta matrix calcualted using all data via the Frobenious norm.

Usage

dynamicsconv(tseries1, tseries2, nitvl, statebounds,lag,
nruns = 10000, CI = 0.95, mineffect = 0.05)

Arguments

tseries1, tseries2	Vectors of values defining two time series. The last observation in tseries1 should be considered to immediately precede the first observation of tseries2. The times at which these values were collected are not recorded, but knowledge of this feature will inform the selection of the lag.
nitvl	Number of evenly spaced intervals to stratify tseries1 and tseries2 into. The first interval uses the last length(tseries1)/nitvl)th of tseries1 and the first length(tseries2)/nitvl)th of tseries2. The second interval uses the last 2*length(tseries1)/nitvl)th of tseries1 and the first 2*length(tseries2)/nitvl)th of tseries2, and so on.
nruns	Number of bootstrap runs to determine significance.
CI	Width of confidence interval to be used to determine significance. Must be within (0,1).
mineffect	Absolute value of smallest effect size that will be considered significant.

Value

A list consisting of five elements:

normdiff	A vector with each element consisting of the norm of the difference in delta matrices between using all of the data and each interval under consideration.
itvl1	Vector of the indices of the intervals used for tseries1. For the ith iteration, indices (itvl1[i]):length(tseries1) were used to form a transition matrix.
itvl2	Vector of the indices of the intervals used for tseries2. For the ith iteration, indices (1:itvl2[i]) were used to form a transition matrix.
ilength1	Length of each interval used for tseries1. It is the cumsum of the differences of itvl1.
ilength2	Length of each interval used for tseries2. It is the cumsum of the differences of itvl2.

vancebee/MarkovSCD documentation built on May 13, 2019, 1:37 a.m.