Classification case: Assessing the performance of remote sensing models

Luciana Nieto & Adrian Correndo

2022-07-24

1. Introduction


The metrica package was developed to visualize and compute the level of agreement between observed ground-truth values and model-derived (e.g., mechanistic or empirical) predicted.

This package is intended to fit into the following workflow:

  1. a data set containing the observed values is used to train a model
  2. the trained model is used to generate predicted
  3. a data frame containing at least the observed and model-predicted values is created
  4. metrica package is used to compute and evaluate the classification model based on observed and predicted values
  5. metrica package is used to visualize model fit and selected fit metrics

This vignette introduces the functionality of the metrica package applied to observed and model-predicted values of a binary land cover classification scenario, where the two classes are vegetation (1) and non-vegetation (0)).

Let’s begin by loading the packages needed.
## Libraries

library(metrica)
library(dplyr)
library(purrr)
library(tidyr)

2. Example datasets

2.1. Kansas Land Cover data



Now we load the binary land_cover data set already included with the metrica package. This data set contains two columns: 

# Load
binary_landCover <- metrica::land_cover

# Printing first observations
head(binary_landCover)
#>   actual predicted
#> 1      0         0
#> 2      1         1
#> 3      1         1
#> 4      0         0
#> 5      0         0
#> 6      1         1

2.2. Maize Phenology