An electronic nose system for the monitoring of water cane shoots quality with swarm clustering algorithm

Abstract

AbstractA fast, portable, nondestructive, and simple‐to‐operate method for determination of water cane shoot quality via an electronic nose (E‐nose) system is presented. The responses of E‐nose sensors to cane shoot samples during storage time are measured. Considering that one of the most common problems with fresh aquatic vegetables is the nearly negligible change in volatile organic compounds, a novel swarm clustering based on particle swarm optimization (SWC) algorithm is proposed to extract the effective features of the sensors responses. In SWC, the number of clusters is automatically estimated based on the neighborhood radius of the particles, which reduces the difficulty in determining the number of cluster. Traditional clustering algorithms, such as K‐means clustering (KMC) and hierarchical clustering analysis (HCA), are used for the purposes of comparison, and the changes in quality of cane shoot samples are quantitatively determined according to the hardness and pH. The results suggest that the E‐nose‐based SWC methodology is capable of accurately classifying the cane shoots freshness during the 20‐day storage time, which is much better than that of the comparative KMC and HCA methods. Additionally，quantitative analysis of the sensory evaluation, hardness, and pH is shown to have good agreement with the E‐nose system analysis results. Validation experiments show that unknown samples are successfully grouped into several different categories with a high correlation (R2 = .982) between unknown and known samples, demonstrating the potential use of E‐nose system to predict the freshness of the cane shoots.