Application of Electronic Nose to Detect Pomegranate Paste Adulteration Using Pattern Recognition Methods and Artificial Neural Network

Adulteration in food products is regarded as a main challenge in food industry, which adversely affects food quality and health. Owing to pleasant taste and antioxidant properties, pomegranate paste is one of the most valuable and desirable foods in the people diets in some countries. As a luxury and expensive food, it is likely to be adulterated by some producers or distributors for the more profits. In this study development and application of machine olfaction system using array of gas sensors to detect adulteration in pomegranate paste was aimed. Principal Component Analysis (PCA), Linear Discrimination analysis (LDA) and Artificial Neural Network (ANN) methods were used to analyze response of the sensor arrays. Based on the results, PCA with two components PC1 and PC2 described 94% of total data variance. In LDA method, the classification accuracy of pomegranate paste samples was 97.65% which higher than PCA method. The values of correlation coefficient (R2) and root mean squared error (RSME) of neural network in ANN method using the structure of 6-9-7 were 0.984 and 0.0018 respectively. This study reveals that the electronic nose device can be used as a non-destructive tool to classify and detect adulteration of different classes of pomegranate paste.