Machine Learning Approach for the Estimation of Optical Properties for a D-Shaped Photonic Crystal Fiber

The field of photonic crystals (PCs) has witnessed significant advancements in recent decades, marked by numerous breakthroughs and innovations for its ability to confine light within a narrow range and manipulate its characteristics. This study focuses on the optical properties of D-shaped photonic crystal fibers (PCFs). It is a prominent area of research due to its compatibility with conventional optical fibers and versatility in supporting various modes. However, accurately determining the optical properties of D-shaped PCFs presents challenges owing to their intricate geometry and material composition. To address this, a machine learning-based model was developed, aiming to predict the optical properties of D-shaped photonic crystal fibers. The dataset of COMSOL Multiphysics was used to train the algorithms. This methodology entails training machine learning models using a dataset of simulated optical properties, enabling accurate predictions for new fibers.