Electronic nose based on gas sensors and a machine-learning algorithm to discriminate potatoes according to the cultivated field nature

Abstract

This article assesses potatoes using an electronic nose according to the nature of the original fields of their harvest: traditionally treated with manure from domestic sheep and donkeys or with manure from chicken farms. A network of five commercial metal oxide sensors, a data card acquisition, a personal computer, and a data analysis and processing approach make up our electronic nose tool. The method of principal component analysis (PCA) was used for the classification of data from both two potatoes kinds and revealed that the first three principal components (PC1, PC2, and PC3) may explain 99.20 percent of the variance by recording a spectacular visual separation allowing each group to be identified.