Hyperparameter optimization of tapping center machines model using robust whale optimization algorithm

To build a synchronization error prediction model for the machine tool efficiently, robust whale optimization algorithm (RWOA) method proposed in this study is applied to hyperparameter of its model. The RWOA integrated non-linear time-invariant inertia weighting (NTIW) and Taguchi-based adaptive parameter exploration (ATPE) improve performance WOA promote robustness. NTIW can other algorithms, so used WOA. In addition, Taguchi get an excellent combination variables with optimal values stable performance, making robust. First, verify validity method, 13 benchmark functions were study. results function tests include mean, standard deviation, p-value t-distribution test. show that 11 differ significantly. non-significant difference functions, means deviations obtained by are considerably better than those Since product cost tools higher, if be built effectively, it reduce cost. Therefore, study, was explore best From results, model’s average MAPE (mean absolute percentage error) 7.2604% training data 9.2603% testing under 30 modeling runs. For one models, 6.8384% 6.7372% data. This also introduced into actual experimental showed 6.3447%. effectively explores suitable machine.