Jayatilaka Sudeepa, Weerasinghe Ruvan, Fernandopulle Neil
The advent of DNA profiling in the recent years has revolutionized law enforcement throughout the world by allowing suspects to be matched with minuscule amounts of biological evidence from a crime scene. It would be of immense importance if the ethnic origin of an unknown DNA sample can be predicted in advance particularly for screening purposes in forensic investigations. To this end population studies have been conducted and it is clear that there exists a correlation between allele frequencies and the ethnicity. This paper presents a neural network solution to forecast ethnicity from DNA profiles. To the best of our knowledge machine learning has not been exercised in this problem domain before.