The Investigation of Possibility of Honey Adulteration Detection Using Support Vector Machine Modeling

Honey have high nutritional value and unique flavor and because of that, the price of natural honey is relatively much higher than hat of other sweeteners. Honey is susceptible to adulteration with cheaper sweeteners; those that have been detected in adulterated honeys include sugar syrups and glucose syrup. In recent years, several methods have been examined to determine honey adulteration. In this study, the Support Vector Machine (SVM) modeling has been used as a way to determine the type and percentage of adulteration in honey. For this reason, glucose syrup and sugar syrup as common sweetener in honey fraud in different percentages of replacement were used and various physicochemical properties of honey were determined. Using SVM, 2 properties (viscosity and color parameters L *) in the detection of counterfeit honey 6 characteristics (non-reducing sugars, sucrose, electrical conductivity, L *, b * and viscosity) in the detection of adulteration and 4 feature (reducing sugars before hydrolysis, reducing sugars after hydrolysis, fructose and refractive index) in the detection of fraud, were selected. The results of the evaluation showed that the models can detect the counterfeit honey than pure honey, the type of fraud and fraud levels in 100% accurat.