Artificial Sensory Systems Combined with UV-Vis Spectrophotometry as a Robust Approach for VOCs Analysis of Human Urine and Exhaled Breath

Abstract

The aim of this work was to investigate the ability of electronic nose (E-nose) and voltammetric electronic tongue (VE-tongue) combined with UV-Vis spectrophotometry to classify breath and urine samples according to the creatinine levels (CLs). Principal Component Analysis (PCA), Support Vector Machines (SVMs) and Partial Least Squares-Regression (PLS-R) were applied for analyzing both E-nose and VE-tongue datasets. As results, PCA on E-nose data represents certain limitations on the discrimination of breath samples depending on three CLs. While SVMs leads to an excellent classification of the studied breath samples. Otherwise, both PCA and SVMs performed on the VE-tongue data revealed that urine samples could be classified accurately according to three CLs. Moreover, PLS-R reveals a good correlation between the VE-tongue and spectroscopic Jaffe's method. In summary, we might say that the electronic sensing systems offer an efficient tools for breath and urine analysis.