Here, MATLAB is used for training the network model for tool life prediction. The network is trained with 70 percent of the measured data. The remaining 15 percent of thedata are used for testing and the other 15 percent of the data are used for validation.The predicted values of neuro fuzzy model are given in Table X. The predicted valuesof the tool life as given by neuro fuzzy model are compared with the experimentalvalues for the validation of experiments. The comparison is depicted in terms ofpercentage error in Figure 11 for validation. It is observed from the Table XI that toollife predicted by the nero-fuzzy model is very close to ±2 percent of the actual values.Most of the values are within the ±2 percent error band.