Change type analysis of the bfast01 function

Description

A function to determine the change type

Usage

1
bfast01classify(object, alpha = 0.05, pct_stable = NULL)

Arguments

object

bfast01 object, i.e. the output of the bfast01 function.

alpha

threshold for significance tests, default 0.05

pct_stable

threshold for segment stability, unit: percent change per unit time (0-100), default NULL

Details

bfast01classify

Value

bfast01classify returns a data.frame with the following elements:

flag_type

Type of shift: (1) monotonic increase, (2) monotonic decrease, (3) monotonic increase (with positive break), (4) monotonic decrease (with negative break), (5) interruption: increase with negative break, (6) interruption: decrease with positive break, (7) reversal: increase to decrease, (8) reversal: decrease to increase

flag_significance

SIGNIFICANCE FLAG: (0) both segments significant (or no break and significant), (1) only first segment significant, (2) only 2nd segment significant, (3) both segments insignificant (or no break and not significant)

flag_pct_stable

STABILITY FLAG: (0) change in both segments is substantial (or no break and substantial), (1) only first segment substantial, (2) only 2nd segment substantial (3) both segments are stable (or no break and stable)

and also significance and percentage of both segments before and after the potentially detected break: "p_segment1", "p_segment2", "pct_segment1", "pct_segment2".

Author(s)

Rogier de Jong, Jan Verbesselt

References

de Jong R, Verbesselt J, Zeileis A, Schaepman M (2013). Shifts in global vegetation activity trends. Remote Sensing, 5, 1117–1133. http://dx.doi.org/10.3390/rs5031117

See Also

bfast01

Examples

1
2
3
4
5
6
library(zoo)
## define a regular time series
ndvi <- as.ts(zoo(som$NDVI.a, som$Time))
## fit variations
bf1 <- bfast01(ndvi)
bfast01classify(bf1, pct_stable = 0.25)