Chemometrics Assisted Method for Classification of Mango Juice as Adulterated or Safe with over Use of Artificial Colours by UV Spectroscopic Data

Tartazine, Sunset Yellow and Beta Carotine are commonly used artificial colours in commercial mango juices in order to make them attractive to the consumers, even though these synthetic colours have hazardous effect on health. Therefore, it is very often necessary to classify juices as adulterated with heavy use of these colours or not. In the present study, two chemometric techniques, Artificial Neural Network (ANN) and Partial Least Square-Discrimination Analysis (PLS-DA) have been assessed for their efficiencies for classification. Here, UV spectroscopic data are used as input. Three techniques, Standard Normal Variate (SNV), Multiplicative Scatter Correction (MSC), Savitzky–Golay (S-G) filtering have been evaluated for their do-noising performance, and select the best one. Before calibration, spectral data are de-noised with MSC as is proved most efficient for de-noising UV spectral data. Spectral range 371-533 nm has been used for calibration ultimately. ANN shows better classification results than PLS-DA for all colours. Finally, the study is proposing a simpler and cheaper method for classification of mango juice as adulterated or safe with over use of artificial colours by applying ANN in de-noised spectroscopic data.