Abstract
Bee pollen is prepared themselves by pollens collecting from plants and has nutritive and therapeutic properties that make it attractive for human health. It has a typical composition related to the botanical origin and geographical location. This study aims to distinguish and identify bee pollen belonging to different Algerian regions and different plants. A methodology for the identification of pollen was developed based on Attenuated Total Reflectance Fourier transform infrared (ATR-FTIR) spectroscopy. This method is simple and fast where samples are not destroyed, also unsupervised statistical methods principal component analysis (PCA) and hierarchical clustering analysis (HCA) are performed. Seventy-two pollen samples were collected and the ATR-FTIR spectra were recorded without processing the samples. ATR-FTIR spectra analysis allowed a reliable determination of the components present in the different samples. Further, PCA and HCA were utilized to evaluate the differences and similarities between the collected samples. Indeed, the PCA score plot and HCA based on ATR-FTIR revealed the same discriminatory trend, where the samples were divided into three main classes based on their total bee pollen. As a result, the PCA along with the HCA was a good and consistent model for identifying and distinguishing pollen grains.