Ensemble regression method for predicting the dependence of electrical resistance on the elongation of polymer composite materials

Abstract Composite materials are actively used as protection against electromagnetic fields, as conductors and to remove excess heat from microcircuits and boards. Such products are often subjected to mechanical stress during operation, and stretched polymer‐film composite materials suffer from deterioration in electrical conductivity, which leads to loss of electrically conductive properties. In this work, the effect of mechanical deformation (stretching) on electrical conductivity properties was studied. A predictive model was built based on Boltzmann statistics, which allows predicting the loss of electrically conductive properties when materials are under mechanical stress. The resulting model was refined by introducing coefficients calculated using an ensemble regression machine learning method. The mean absolute percentage error of the predicted electrical conductivity value for the materials under mechanical stress is about 5%–20%. Highlights Stretching influence on conductivity and shielding. An expression for description and prediction resistance during stretching. Machine learning for predictive modeling.