evaluateCycle: Evaluation Cycle of EOT-Based NDVI Resampling

In environmentalinformatics-marburg/ESD:

Description

Evaluate the performance of EOT-based spatial resampling based on selected error and regression metrics, including mean error (ME), mean absolute error (MAE), root-mean-square-error (RMSE) as well as correlation coefficient (r) and coefficient of determination (R^2).

Usage

evaluateCycle(pred, resp, training, n = 10L, var = NULL,
  deseasoning = FALSE, cores = 1L, ...)

Arguments

pred	Raster* object. Predictor data for EOT analysis.
resp	Raster* object. Response data for EOT analysis.
training	integer. Indices of training layers in 'pred' and 'resp'. All non-included layers are automatically used for model testing.
n	integer. Number of EOT modes to calculate.
var	numeric. Minimum amount of variance to be explained. If NULL (default), all calculated EOT modes are used for predict-ion.
deseasoning	logical. If TRUE, deseason-ing is applied prior to the actual EOT procedure.
cores	integer. Number of cores for parallel processing.
...	Additional arguments passed to eot.

Value

A list of length 2. The content of the first (second) slot is a data.frame (RasterLayer) with spatial (temporal) error and regression metrics (RMSE).

Author(s)

Florian Detsch

References

Detsch F, Otte I, Appelhans T, Hemp A, Nauss T (2016) Seasonal and long-term vegetation dynamics from 1-km GIMMS-based NDVI time series at Mt. Kilimanjaro, Tanzania. Remote Sensing of Environment 178, 70-83, doi:10.1016/j.rse.2016.03.007.

See Also

evaluate, deseason, eot, nXplain.

environmentalinformatics-marburg/ESD documentation built on May 16, 2019, 7:49 a.m.