is very popular milk product used to make variety of sweets in India.
is made by milk thickening and heating it in an open iron pan. In this study, feedforward Backpropagation Neural Network (BPNN), Radial Basis Function Neural Network (RBFNN) and Multiple Linear Regression (MLR) models have been developed to predict shelf life of cow milk
stored at 37
C. Five input parameters,
., moisture, titratable acidity, free fatty acids, tyrosine and peroxide value are considered to predict sensory score. The dataset comprised of 48 observations. The accuracy of these models was judged with percent Root Mean Square Error (%RMSE). The BPNN model with Bayesian regularization algorithm provided static and consistent results. The residual shelf life of
was also computed using regression equations based on sensory scores. The BPNN model exhibited the best fit (%RMSE, 4.38) followed by MLR model (%RMSE, 9.27) and RBFNN model (%RMSE, 10.84).