Use of intellectual technologies for quality control of curd

Method of curds quality control, based on neural network model is introduced in the paper. It allows to give fair estimate of final product’s quality without use of professional degustators. Method consists in construction of neural network model to estimate quality of curds by input and output parameters. Hardware software complex of curds taste control is proposed to apply in industry for independent estimate of final product. Operation of neural network model based on usage of artificial neural networks, which is one of directions of artificial intelligence theory. Neural network model consists of several artificial neurons layers, which emulate nerve cells functioning. Feed forward network of MLP type was used in solving of curds taste control problem. Feature of this network is that signals passed from one layer’s neurons to next layer’s neurons only from input layer to output, not the other way. Neural network was trained for its correct operation by selection of optimal synaptic factors. Flow diagram of training algorithm is introduced in the paper. Algorithm of neural network tuning is described in the paper. Author proposed set of recommendations for software-hardware complex deployment. Experiment results show that this taste estimate method based on neural net works simplify a problem of final product control as allow to react to deviations in production process on proper time.