The use of support vector machine, neural network, and regression analysis to predict and optimize surface roughness and cutting forces in milling

Abstract

In the present study, prediction and optimization of the surface roughness and cutting forces in slot milling of aluminum alloy 7075-T6 were pursued by taking advantage of regression analysis, support vector regression (SVR), artificial neural network (ANN), and multi-objective genetic algorithm. The effects of process parameters, including cutting speed, feed per tooth, depth of cut, and tool type, on the responses were investigated by the analysis of variance (ANOVA). Grid search and cross-validation methods were used for hyperparameter tuning and to find the best ANN and SVR models. The training algorithm of developed NNs was one of the hyperparameters which was chosen from Levenberg-Marquardt and RMSprop algorithms. The performance of regression, SVR, and ANN models were compared with each other corresponding to each machining response studied. The ANN models were integrated with the non-dominated sorting genetic algorithm (NSGA-II) to find the optimum solutions by means of minimizing the surface roughness and cutting forces. In addition, the desirability function approach was utilized to select proper solutions from the statistical tools.