Package 'LactCurveModels'

LactCurveModels: Lactation Curve Model Fitting for Dairy Animals

Description

Fits up to 20 nonlinear lactation curve models to daily milk yield data from one or more dairy animals. For each animal the package produces:

• Four CSV tables (parameter estimates, actual vs predicted values, residuals, and a goodness-of-fit summary).

• Fifteen diagnostic figures (model overlays, ranked panels, residual diagnostics, bubble charts, bar charts for R-squared (R2), Root Mean Square Error (RMSE), Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), Durbin-Watson (DW) statistic, correlation heatmap, and peak-yield charts).

When multiple animals are analysed, three additional combined cross-animal outputs are produced: an R2 comparison figure, an RMSE comparison figure, and a best-model overlay figure, plus combined CSV tables.

Quick start

Note: out_dir must always be supplied — choose any folder on your system where you want results saved.

library(LactCurveModels)

# All animals, all 20 models
results <- run_lactation_analysis(
  input_csv_path = "path/to/mydata.csv",
  out_dir        = "path/to/results"
)

# Selected animals and models
results <- run_lactation_analysis(
  input_csv_path   = "path/to/mydata.csv",
  selected_models  = c("Wood_1967", "Wilmink_k005", "Ali_Schaeffer"),
  selected_animals = c("Animal_A", "Animal_B"),
  out_dir          = "path/to/results"
)

Input CSV format

The CSV file must contain exactly three columns:

animal_id

Character identifier for each animal.

dim

Integer fortnightly time index (1 to 20).

dmy

Numeric daily milk yield (kg/day).

Author(s)

Maintainer: Raja TV [email protected]

Authors:

• Lalremruati PC [email protected]

• Priyadharshini P [email protected]

• Nidhishree NS [email protected]

• Dheeraj Gurjar [email protected]

• Rani Alex [email protected]

• Vikas Vohra [email protected]

See Also

Useful links:

• https://github.com/venkatesanraja/LactCurveModels

• Report bugs at https://github.com/venkatesanraja/LactCurveModels/issues

---

Fit lactation curve models to a single animal's data

Description

Attempts to fit up to 20 nonlinear lactation curve models using nls() with the "port" algorithm. Models that fail to converge are silently skipped.

Usage

fit_lactation_models(data, selected_models = "all")

Arguments

data

A data frame with columns x (time index), y (milk yield kg/day), and z (x/365, for Morant-Gnanasakthy). Create it with make_animal_df().

selected_models

Either "all" (default) to fit all 20 models, or a character vector of model names to fit a subset. Valid names: "Brody_1923", "Brody_1924", "Parabolic_Exp_Sikka", "Wood_1967", "Wood_Dhanoa", "Cobby_LeDu", "Quadratic_Dave", "Mixed_Log_GS", "Khandekar", "Wilmink_k005", "Wilmink_k_estimated", "Ali_Schaeffer", "Cappio_Borlino", "Papajcsik_Bodero1", "Papajcsik_Bodero2", "Papajcsik_Bodero3", "Inverse_Poly_Nelder", "Log_Quadratic_Adediran", "Morant_Gnanasakthy", "Pollott_Multiplicative".

Value

A named list. Each element corresponds to one successfully fitted model and contains: model (nls object), metrics (list of fit statistics), predictions (fitted values), std_errors (parameter SEs).

Examples

# Minimal toy example (auto-tested)
toy <- data.frame(
  x = 1:10,
  y = c(8, 11, 13, 12, 11, 10, 9, 8, 7, 6),
  z = (1:10) / 365
)
fits <- fit_lactation_models(toy, selected_models = "Wood_1967")


# Full run with all 20 models (takes longer)
fits_all <- fit_lactation_models(toy)

# Selected models only
fits_sub <- fit_lactation_models(toy,
              selected_models = c("Wood_1967", "Wilmink_k005"))

---

Run lactation curve analysis for one or more dairy animals

Description

Reads a CSV file containing milk yield records, fits up to 20 nonlinear lactation curve models for each animal (or a selected subset of animals and/or models), saves per-animal diagnostic figures and CSV tables, and produces combined cross-animal summary outputs.

Usage

run_lactation_analysis(
  input_csv_path,
  selected_models = "all",
  selected_animals = "all",
  out_dir
)

Arguments

input_csv_path	Character. Full path to the input CSV file. The file must contain exactly these three columns (any column order): animal_id Character or factor identifying each animal. dim Integer or numeric time index (days-in-milk fortnightly index, i.e. 1, 2, ..., 20). dmy Numeric daily milk yield (kg/day).
selected_models	Either "all" (default) to fit all 20 lactation curve models, or a character vector of model names. Valid names: "Brody_1923", "Brody_1924", "Parabolic_Exp_Sikka", "Wood_1967", "Wood_Dhanoa", "Cobby_LeDu", "Quadratic_Dave", "Mixed_Log_GS", "Khandekar", "Wilmink_k005", "Wilmink_k_estimated", "Ali_Schaeffer", "Cappio_Borlino", "Papajcsik_Bodero1", "Papajcsik_Bodero2", "Papajcsik_Bodero3", "Inverse_Poly_Nelder", "Log_Quadratic_Adediran", "Morant_Gnanasakthy", "Pollott_Multiplicative".
selected_animals	Either "all" (default) to analyse every animal present in the CSV, or a character vector of animal IDs that match entries in the animal_id column exactly.
out_dir	Character. Base output directory. A sub-folder is created for each animal. Must be supplied explicitly — there is no default. The directory is created automatically if it does not already exist. Example: out_dir = "D:/lactation_results".

Value

A named list (one element per animal) each containing:

id

Animal ID string.

model_fits

Named list of fitted model objects and metrics.

param_table

Parameter estimates data frame.

pred_table

Actual vs predicted data frame.

resid_table

Residuals data frame.

metrics_df

Per-model goodness-of-fit metrics, sorted by R2.

summary_metrics

Wide summary metrics data frame.

Returned invisibly so large objects are not auto-printed.

Examples

# Build a small example CSV in a temporary directory
tmp_csv <- file.path(tempdir(), "example_animals.csv")
example_data <- data.frame(
  animal_id = rep(c("Animal_A", "Animal_B"), each = 20),
  dim       = rep(1:20, times = 2),
  dmy       = c(
    8.1, 11.2, 13.0, 12.5, 11.8, 11.0, 10.4, 9.8, 9.3, 8.8,
    8.3, 7.9, 7.5, 7.1, 6.8, 6.5, 6.2, 5.9, 5.7, 5.4,
    7.5, 10.1, 11.8, 11.3, 10.7, 10.0, 9.4, 8.9, 8.4, 7.9,
    7.5, 7.1, 6.8, 6.4, 6.1, 5.8, 5.6, 5.3, 5.1, 4.9
  )
)
write.csv(example_data, tmp_csv, row.names = FALSE)

# Analyse all animals with two selected models
results <- run_lactation_analysis(
  input_csv_path  = tmp_csv,
  selected_models = c("Wood_1967", "Wilmink_k005"),
  out_dir         = tempdir()
)

# Analyse one animal only
results2 <- run_lactation_analysis(
  input_csv_path   = tmp_csv,
  selected_models  = c("Wood_1967", "Ali_Schaeffer"),
  selected_animals = "Animal_A",
  out_dir          = tempdir()
)

---