Study on Single and Double Hidden Layers of Cascade Artificial Neural Intelligence Neurocomputing Models for Predicting Sensory Quality of Roasted Coffee Flavoured Sterilized Drink

For centuries, coffee has been brewed and consumed in households, hot shops and restaurants. Today flavoured milks have become very popular and they contain nutrients as compared with soft drinks. Sterilized milk is the product made by heating milk to high temperature (121 C) with 15 m holding time so that it remains fit for human consumption for longer time at room temperature. Efficiency of single and double hidden layers of Cascade neurocomputing models for prediction of sensory quality of roasted coffee flavoured sterilized drink were studied. Colour and appearance, viscosity, flavour and sediment were taken as input parameters, while overall acceptability was used as output parameter. The results of cascade neurocomputing models were calculated with two types of prediction performance measures, viz., root mean square error and coefficient of determination R.The study revealed that more the number of neurons in single hidden layer, less the error for cascade neurocomputing models ( RMSE:0.00011; R : 0.999999; neurons:50). General Terms Artificial Neural Network, MATLAB, Soft Computing