Abstract
This research introduces a methodology for data-driven regression modeling of components exhibiting nonlinear characteristics, utilizing the sparse identification of nonlinear dynamics (SINDy) method. The SINDy method is extended to formulate regression models for interconnecting components with nonlinear traits, yielding governing equations with physically interpretable solutions. The proposed methodology focuses on extracting a model that balances accuracy and sparsity among various regression models. In this process, a comprehensive model was generated using linear term weights and an error histogram. The applicability of the proposed approach is demonstrated through a case study involving a sponge gasket with nonlinear characteristics. By contrasting the predictive model with experimental responses, the reliability of the methodology is verified. The results highlight that the regression model, based on the proposed technique, can effectively establish an accurate dynamical system model, accounting for realistic conditions.