The use of 3D-nonlinear regression analysis in mathematics modeling of colour change in roasted hazelnuts

In this research, round hazelnut cultivars (Foşa, Palaz and Tombul) harvested from the Ordu, Giresun and Trabzon regions in 2001 and 2002 were supplied by the Fiskobirlik Company. Changes in L, a and b values of hazelnut varieties during different roasting applications (125–165 °C, 10–55 min) were used to determine the optimum conditions of the roasting process and develop mathematical equations. As a result of variance analysis, the effects of the roasting process, cultivars and measurement types on colour values (L, a and b) of hazelnuts were found to be statistically significant (p < 0.01). Values of a and b for hazelnut samples showed an increase while L decreased. The regression coefficients (R2) which described the prediction models for the dependent variables ranged from 0.73 to 0.97. In addition, changes in colour values, particularly a and L values, gave high R2 and the obtained three-dimensional (3D) nonlinear equations may be utilized to determine the optimum roasting degree based on time and temperature.