Tüketicilerin Gıda Tercihlerine Yardımcı Olmak Üzere Yakın Kızılötesi Spektroskopisi’nin Kullanılması

Abstract

In sustainable food systems, innovations using new technology are needed to facilitate consumers' food choices. Near infrared spectroscopy is a spectroscopic method used in the analysis of organic compounds of foods, determination of counterfeit and adulteration in foods, and determination of protein, fat, dry matter and composition properties. The integration of this technology, which is widely used in the food industry, into the field is very important for consumers to make their product choices more easily and reliably. From this point of view, it is aimed to develop an artificial intelligence model that can decide whether the food products they want to buy in online shopping applications and / or local markets on behalf of consumers are suitable for them using YKS technology. In this context, white cheese was determined as the study group due to its widespread use. The spectra of compositional properties of 15 different feta cheese samples selected from local food markets in Turkey were taken using YKS technology. Then, sensory analysis of all cheeses was made with randomly selected consumers (n=100) and cheeses were evaluated in terms of taste, smell, elasticity, hardness, color and general taste parameters using emoji. At the same time, the determination of fat, protein, dry matter, salt content and acidity ratios, and texture and color analyzes of cheeses were made. A data set was created using the SPSS data analysis system with all the obtained data and the relationship between the parameters was determined. In addition, with the results of sensory analysis of YKS spectra using principal component analysis (PCA), one of the data analysis systems; A significant correlation with the results of chemical and physical analysis was performed using partial small squares regression (PLS). With the chemometric study, feta cheese samples could be grouped according to their spectral, compositional and sensory properties, and in this way, a estimation model could be developed in which the consumer, who introduced the sensory criteria to the system by tasting, could evaluate whether the product with the product properties information was suitable for his/her sensory criteria. It is predicted that this developed forecasting model will be a marketing model with the potential to serve the product development of the food industry in the future and to help consumers make informed choices.