Foundations

Forensic Anthropologist Goals

Include locating and recovering remains, estimating a biological profile, analyzing skeletal trauma, and observing taphonomy, with applications in justice, human rights, and disaster victim identifications.

Daubert Guidelines

Established rules for evidence admissibility and standardization, leading forensic anthropology to standardize methods for court admissibility.

Reliability

Refers to getting consistent results upon repetition, ensuring the method's repeatability.

Validity

Indicates the likelihood of being correct more often than by chance, with internal and external validity types.

Error Sources

Include practitioner, instrument, statistical, and method errors, affecting the accuracy of forensic anthropology methods.

Supervised Machine Learning

Utilized for labeled data, where the model learns from training data sets to make predictions in classifications and regressions.

Unsupervised Machine Learning

Applied to unlabeled data to discover natural data groupings without assumptions, including cluster analyses and dimensionality reductions.

Power Analysis

Determines the minimum sample size for quantitative research, emphasizing the importance of sample diversity for model training.

Model Evaluation

Involves training and testing data to assess the model's performance, with training error and testing error capturing internal and external validity.

Validation Methods

Include hold-out/independent test set, K-fold, LOO, and bootstrap, each suitable for different sample sizes to evaluate model performance.

Sources of error

practitioner error, instrument error, statistical error and method error