Extreme learning machine neural networks for adult skeletal age-at-death estimation

Abstract The aim of this chapter is to introduce the extreme learning machine algorithm to construct single-layer feedforward neural networks for age-at-death estimation. The most relevant notation and machine learning concepts are approached in context focusing on the most efficient algorithms to train this type of neural network. Conformal prediction, a technique used to construct confidence measures for machine learning algorithms, is introduced to derive prediction intervals for machine-based age-at-death estimation. A performance analysis is conducted to assess different aspects of this algorithm and how it performs on the difficult task of adult age-at-death estimation from skeletal remains.