This function computes the Residual Prediction Deviation (RPD), which
is defined as the standard deviation of observed values divided by
the Root Mean Square Error or Prediction (RMSEP). The RDP takes both
the prediction error and the variation of observed values into 
account, providing a metric of model validity that is more objective
than the RMSEP and more easily comparable across model validation
studies. The greater the RPD, the better the model's predictive
capacity.

model

Validation

The phenology of plants (i.e. the timing of their annual life
phases) depends on climatic cues. For temperate trees and many other plants,
spring phases, such as leaf emergence and flowering, have been found to result
from the effects of both cool (chilling) conditions and heat. Fruit tree
scientists (pomologists) have developed some metrics to quantify chilling
and heat (e.g. see Luedeling (2012) <doi:10.1016/j.scienta.2012.07.011>).
'chillR' contains functions for processing temperature records into
chilling (Chilling Hours, Utah Chill Units and Chill Portions) and heat units
(Growing Degree Hours). Regarding chilling metrics, Chill Portions are often
considered the most promising, but they are difficult to calculate. This package
makes it easy. 'chillR' also contains procedures for conducting a PLS analysis
relating phenological dates (e.g. bloom dates) to either mean temperatures or
mean chill and heat accumulation rates, based on long-term weather and phenology
records (Luedeling and Gassner (2012) <doi:10.1016/j.agrformet.2011.10.020>).
As of version 0.65, it also includes functions for generating weather
scenarios with a weather generator, for conducting climate change analyses
for temperature-based climatic metrics and for plotting results from such
analyses. Since version 0.70, 'chillR' contains a function for interpolating
hourly temperature records.

Eike Luedeling

chillR

Statistical Methods for Phenology Analysis in Temperate Fruit
Trees

Lars Caspersen

Eduardo Fernandez

RPD function

<dl><dt>predicted</dt>
<dd>a numeric vector containing predicted values.</dd>
<dt>observed</dt>
<dd>a numeric vector of the same length as ```predicted```
containing observed values.</dd>
<dt>na.rm</dt>
<dd>Boolean parameter indicating whether NA values should be removed before the analysis</dd></dl>

Arguments

Author

Residual Prediction Deviation (RPD) — RPD

<dl>

<dt>predicted</dt>
<dd>a numeric vector containing predicted values.</dd>


<dt>observed</dt>
<dd>a numeric vector of the same length as ```predicted```
containing observed values.</dd>


<dt>na.rm</dt>
<dd>Boolean parameter indicating whether NA values should be removed before the analysis</dd>

</dl>

RPD: Residual Prediction Deviation (RPD)

Description

Usage

Value

Arguments

Author

Details

References

Examples