Optical Properties Prediction of Negative Dispersion-Compensating Photonic Crystal Fiber Using Machine Learning

Photonic Crystal Fibers (PCFs) are used in spectroscopy, imaging, and metrology as well as in long-haul optical communication systems for their dispersion compensating characteristics. In this work, a novel and highly negative dispersion compensating photonic crystal fiber is structured, and then the study of machine learning approaches has been proposed to predict the output properties like effective refractive index, dispersion, confinement loss, effective area, and V-parameter from input parameters in the range of wavelength from 1.18–1.75$\mu$m, pitch from 0.75–0.9$\mu$m, the diameter of the core, and air holes in the cladding region. The proposed models take fewer computing resources and less time than COMSOL Multiphysics simulation and Artificial Neural Network. The machine learning models take milliseconds to train and less than one millisecond to test. The proposed PCF with negative dispersion characteristics has the potential for applicability in real high-rate optical communication. In hence, machine learning approaches are considered as an alternative to conventional numerical simulation to predict optical properties.