Akademik Veri Yönetim Sistemi
Final SensorFINT International Conference, Cordoba, İspanya, 29 - 31 Mayıs 2024, ss.42, (Özet Bildiri)
Sucrose is one of the main ingredients used in sugar industry. Worldwide sucrose is produced from two different sources, sugar beet and sugar cane. However, different sources have been dominating the market due to governmental policies. So, it is important to understand the source of the sugar in the market. Since molecular structure of sucrose is same, it is difficult to differentiate sources by using chemical methods. Therefore, developing more practical and affordable methods would be valuable for food industry. NIR spectroscopy is a promising technique that can detect the differences in the plant base and production steps in a non-destructive manner (Morellos et al., 2016). In this research, sucrose samples from different plant sources (cane, beet) were collected from 12 countries to prepare crystal sucrose samples and their absorbances were recorded between 1000-2500 nm. Measurements were directly taken from sucrose crystals as they were collected from local markets. Results showed that in all regions spectral signature differences were observable. First a PCA with 2 components was conducted to have an exploratory analysis then prediction algorithms were used. It is known that improving the prediction accuracy requires chemometric approaches such as spectral preprocessing and k-Nearest-Neighbor (kNN) (Fetitah et al., 2021). The results indicate that several methods showed high performance, but kNN gave 97% correct classification with 10 fold cross-validation. The obtained results seemed promising that the plant source of sucrose can be detected by using NIR region and chemometric methods.