Age at death estimation in adult skeletons is hampered, among others, by the unremarkable correlation of bone estimators with chronological age, implementation of inappropriate statistical techniques, observer error and skeletal incompleteness or destruction. Therefore, it is beneficial to consider alternative methods to assess age at death in adult skeletons. The decrease of bone mineral density with age was explored to generate a method to assess age at death in human remains. A connectionist computational approach, artificial neural networks, was employed to model femur densitometry data gathered in 100 female individuals from the Coimbra Identified Skeletal Collection. Bone mineral density declines consistently with age and the method performs appropriately, with mean absolute differences between known and predicted age ranging from 9.19 to 13.49 years. The proposed method – DXAGE – was implemented online to streamline age estimation. This preliminary study highlights the value of densitometry to assess age at death in human remains.
Key-words: proximal femur; BMD; age-at-death estimation; neural networks; forensic anthropology
Please read our paper, before using the app DXAGE: A New Method for Age at Death Estimation Based on Femoral Bone Mineral Density and Artificial Neural Networks.
David Navega, João d'Oliveira Coelho, Eugénia Cunha, Francisco Curate
MAE - mean absolute error; RMAE - relative mean absolute error; MAPE - mean absolute percentage error; RMSE - root mean squared error; RRMSE - relative root mean squared error; RSquared - R-Squared; AdjRSquared - Adjusted R-Squared; PIWidth - Predicted Interval Mean Width; %C - Coverage.