rinform (version 1.0.2)

integration_evidence: Evidence of Integration

Description

Given a sequence of `n` observed states of `l` random variables (series), compute the evidence of integration for each partitioning of the `l` variables, and return the minimum and maximum evidence for each observation. If the number of variables `l` is large, the user can test a small subset of the partitioning schemata by providing a partitioning parts to increase confidence that the system is integrated. In this case, the function computes and returns the evidence of integration for each observation with respect to the partitioning parts.

Usage

integration_evidence(series, parts = NULL)

Arguments

series

Matrix specifying two or more time series.

parts

Vector giving a specific partitioning to use.

Value

A list containing minimum, min, and maximum, max, values of integration for each observation and logicals indicating if the system is integrated or, if parts is provided, returns a vector giving the evidence of integration for each observations.

Examples

# NOT RUN {
# Evidence of Integration of three time series:
series      <- matrix(0, nrow = 10, ncol = 3)
series[, 1] <- c(0, 1, 0, 1, 1, 1, 0, 0, 1, 0)
series[, 2] <- c(0, 1, 0, 1, 1, 1, 0, 0, 1, 0)
series[, 3] <- c(1, 1, 1, 1, 1, 0, 0, 0, 0, 0)
# min        = c(-0.322 0.263 -0.322 0.263 0.263 -0.322 0.263 0.263 -0.322 0.263)
# max        = c( 1.000 1.263  1.000 1.263 1.263  1.000 1.263 1.263  1.000 1.263)
# integrated = c( FALSE  TRUE  FALSE  TRUE  TRUE  FALSE  TRUE  TRUE  FALSE  TRUE)
integration_evidence(series)

# Evidence of Integration of three time series for the partitionning c(1, 1, 2)
parts <- c(1, 1, 2)
# -0.322 0.263 -0.322 0.263 0.263 -0.322 0.263 0.263 -0.322 0.263
integration_evidence(series, parts)
# }