Human Skin Scent: Class and Individual Identification

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F60461373%3A22340%2F22%3A43924417" target="_blank" >RIV/60461373:22340/22:43924417 - isvavai.cz</a>

Výsledek na webu

—

DOI - Digital Object Identifier

—

Jazyk výsledku

angličtina

Název v původním jazyce

Human Skin Scent: Class and Individual Identification

Popis výsledku v původním jazyce

In this work, a total of 504 human scent samples of 40 people (20 women and 20 men) were taken and analyzed using comprehensive two-dimensional gas chromatography with mass spectrometry (GC×GC-MS). The aim of this work was to find trends in the representation of monitored compounds in human scents in connection with sex and to create classification models, which were able to correctly assign sex with the greatest possible probability (class identification). A total of 70 pre-selected compounds were monitored. Various multi-dimensional methods were used for this purpose, namely principal component analysis (PCA), orthogonal partial least squares discrimination analysis (OPLS-DA), quadratic discriminant analysis (QDA), and the supporting vector machine (SVM). In addition, classification models were subsequently sought, which would be able to assign the scent sample to a specific volunteer with the greatest possible accuracy (individual identification). Models based on SVM with a polynomial kernel function were ranked best. Within the framework of sex differentiation, the model created from all (504) measured scent samples achieved a validation accuracy of over 91%, and the SVM model for individual identification achieved a validation accuracy of over 73 %

Název v anglickém jazyce

Human Skin Scent: Class and Individual Identification

Popis výsledku anglicky

In this work, a total of 504 human scent samples of 40 people (20 women and 20 men) were taken and analyzed using comprehensive two-dimensional gas chromatography with mass spectrometry (GC×GC-MS). The aim of this work was to find trends in the representation of monitored compounds in human scents in connection with sex and to create classification models, which were able to correctly assign sex with the greatest possible probability (class identification). A total of 70 pre-selected compounds were monitored. Various multi-dimensional methods were used for this purpose, namely principal component analysis (PCA), orthogonal partial least squares discrimination analysis (OPLS-DA), quadratic discriminant analysis (QDA), and the supporting vector machine (SVM). In addition, classification models were subsequently sought, which would be able to assign the scent sample to a specific volunteer with the greatest possible accuracy (individual identification). Models based on SVM with a polynomial kernel function were ranked best. Within the framework of sex differentiation, the model created from all (504) measured scent samples achieved a validation accuracy of over 91%, and the SVM model for individual identification achieved a validation accuracy of over 73 %

Druh

O - Ostatní výsledky

CEP obor

—

OECD FORD obor

10406 - Analytical chemistry

Projekt

<a href="/cs/project/VJ01030005" target="_blank" >VJ01030005: Vytvoření mezinárodní komunity v oboru „Forenzní olfaktronika“</a><br>

Návaznosti

P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Rok uplatnění

2022

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů