Please use this identifier to cite or link to this item:
http://hdl.handle.net/10609/90946
Title: | Implementación de algoritmos de machine learning para la identificación de relaciones familiares e identificación de desaparecidos mediante STRs de ADN autosómico |
Author: | Luque Gutiérrez, Juan Antonio |
Tutor: | Vegas Lozano, Esteban |
Others: | Sánchez-Pla, Alex |
Abstract: | The goal of this TFM has been to implement and evaluate several machine learning algorithms in order to determine the kinship between two individuals in the context of Disaster Victim Identification (DVI). The development has been done with R and Rstudio. Following the mendelian transmission, the DNA profiles of autosomal Short Tandem Repeat (STR) markers were generated for a group of extensive synthetic families. The different machine learning models were trained and validated using as input data pairs of individuals selected from these synthetic families, with diverse kinships. After evaluating several models, a keras/tensorflow neural network was implemented and trained so that it could be used to predict the relationship of two individuals given the multiple DNA profiles obtained in a DVI. Ten settings of events with multiple victims (from 6 to 200 victims) were simulated, with different levels of difficulty. The inference of the neural network in such settings has solved most of the identifications in a few minutes. In some of the settings, the identifications were solved with a 100% accuracy, while the most complicated setting with the farthest family relationships could only obtain a 50% of accuracy. |
Keywords: | machine learning DNA missing persons identification |
Document type: | info:eu-repo/semantics/masterThesis |
Issue Date: | Feb-2019 |
Publication license: | http://creativecommons.org/licenses/by-nc-nd/3.0/es/ |
Appears in Collections: | Trabajos finales de carrera, trabajos de investigación, etc. |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
jaluquegTFM0219memoria.pdf | Memoria del TFM | 4,87 MB | Adobe PDF | View/Open |
Share:
This item is licensed under a Creative Commons License