author:
- Chi Mei Sun
A Bayesian imputation approach that uses behavioral information captured by an automated nest system and hand-counted egg numbers to estimate individual egg numbers under uncertain identification. An R package for analyzing chicken egg laying patterns, imputing missing laying dates using Hidden Markov Models (HMM), and assigning eggs to animals using probabilistic methods.
- Data Preparation: Clean and format raw autonest metadata
- Quality Control: Flagging of suspicious autonest records
- Bayesian Imputation: A probabilistic model that iteratively updates prior probabilities for egg-hen pairs using behavioral likelihoods
- Egg Assignment: Generate Probabilistic or deterministic assignment of eggs to animals
- Validation: Cross-validation tools for model evaluation
Egg laying behavior is modeled through three complementary behavioral scenarios:
- Fitted using Hidden Markov Models (HMM) to identify laying states
- Implementation:
package
- Recent nest selection patterns (± ?days)
- Penalizes frequent nest switching
- Historical nest usage patterns
# Install devtools if needed
install.packages("devtools")
# Install EggImpute
devtools::install_github("ChiMeiSun/EggImpute")
# Load the package
library(EggImpute)