Machine Learning Approach to Predict Dielectric Permittivity of PE/TiO2 Nanocomposites

Abstract

Controlling process parameters has significant influence in designing and developing nanocomposites materials with tailored dielectric properties. In the present study, polyethylene/TiO2 nanocomposites were fabricated using ball milling technique. The effects of TiO2 nanoparticles on the final dielectric properties of the nanocomposites in frequency domain were investigated. The dielectric spectroscopy measurements revealed that relative dielectric permittivity of the nanocompsoites was increased with TiO2 content. Besides, machine learning approach based on artificial neural networks (ANNs) algorithm was used to predict the dielectric permittivity of the nanocomposites materials. Modeling results showed clearly that the predicted data of the proposed artificial model are in good agreement with the experimental values. Moreover, the present study proved that ANNs can be used as successful tool to predict the dielectric properties of nanocomposites materials.