metrics {OenoKPM}

Performs the modeling of the observed data and returns the fit metrics of the studied model

Description

A function that, based on the observed data, the independent variable (e.g. time in h) and the dependent variable (e.g. CO₂ production in g L^-1), performs the modeling of the fermentation curve based on the chosen model (5PL, Gompertz, or 4PL) and returns the model fit metrics.

As a result, the fit metrics for the chosen model are returned in the form of data.frame: Correlation, R²,Residual sum of squares (RSS_min) and Residual standard error.

Usage

metrics(
  data,
  model,
  save.xls = FALSE,
  dir.save,
  xls.name,
  startA,
  startB,
  startC,
  startD,
  startG
)

Arguments

data

Data frame to be analyzed. The data frame must be in the following order:

• First: All columns containing the independent variable (e.g. time in hours)

• Second: All columns containing dependent variables (e.g. CO₂ g L^-1 production)

• Header: Columns must contain a header. If the treatment ID is in the header, this ID will be used to identify the metrics for each analyzed curve.

model

Model to be adjusted. Argument for model:

• Model = 1. 5PL Model (five-parameter logistic (5PL) model)

• Model = 2. Gompertz Model

• Model = 3. 4PL Model (four-parameter logistic (4PL) model)

save.xls

If TRUE, an xlsx file containing the metrics will be saved in the working directory. If FALSE, the xlsx file will not be saved.

dir.save

Directory path where the xlsx file is to be saved.

xls.name

File name. Must contain the format. For example, "Metrics.xlsx".

startA

Starting estimate of the value of A for model.

startB

Starting estimate of the value of B for model.

startC

Starting estimate of the value of C for model.

startD

Starting estimate of the value of D for model.

startG

Starting estimate of the value of G for model.

Details

Curve fitting from the observed data is performed by the nlsLM() function in the 'minpack.lm' package.

Value

The metrics from the analyzed model are returned in a data.frame. In addition, a "Metrics.xlsx" file can be generated, containing the model fit metrics for each fermentation curve studied: Correlation; R²; Residual standard error; Residual sum of squares (RSS_min).

Author(s)

Angelo Gava

Examples

#Creating a data.frame. 
#First, columns containing independent variable.
#Second, columns containing dependent variable.
#The data frame created presents two 
#fermentation curves for two yeasts with 
#different times and carbon dioxide production.

df <- data.frame('Time_Yeast_A' = seq(0,280, by=6.23),
                 'Time_Yeast_B' = seq(0,170, by=3.7777778),
                 'CO2_Production_Yeast_A' = c(0,0.97,4.04,9.62,13.44,17.50,
                                              24.03,27.46,33.75,36.40,40.80,
                                             44.24,48.01,50.85,54.85,57.51,
                                             61.73,65.43,66.50,72.41,75.47,
                                             77.22,78.49,79.26,80.31,81.04,
                                             81.89,82.28,82.56,83.13,83.62,
                                             84.11,84.47,85.02,85.31,85.61,
                                             86.05,86.27,85.29,86.81,86.94,
                                             87.13,87.33,87.45,87.85),
                 'CO2_Production_Yeast_B' = c(0,0.41,0.70,3.05,15.61,18.41,
                                              21.37,23.23,28.28,41.28,43.98,
                                              49.54,54.43,60.40,63.75,69.29,
                                              76.54,78.38,80.91,83.72,84.66,
                                              85.39,85.81,86.92,87.38,87.61,
                                              88.38,88.57,88.72,88.82,89.22,
                                              89.32,89.52,89.71,89.92,90.11,
                                              90.31,90.50,90.70,90.90,91.09,
                                              91.29,91.49,91.68,91.88))

#Using the metrics() function to find the 
#model fit metrics

metrics(data = df,
model = 1, 
startA = 0,
startB = 1.5,
startC = 500,
startD = 92, 
startG = 1500,
save.xls = FALSE) #5PL Model adopted

metrics(data = df,
model = 2,
startA = 92,
startB = 1.5,
startC = 0,
startD = NA, 
startG = NA, 
save.xls = FALSE) #Gompertz Model adopted

metrics(data = df,
model = 3,
startA = 0,
startB = 2.5,
startC = 10,
startD = 92, 
startG = NA, 
save.xls = FALSE) #4PL Model adopted 

#Saving an xlsx file. In this example, 
#we will use saving a temporary file in 
#the temporary file directories.