Modeling and realization of photonic biosensor for hazardous virus detection using ML approach

The broad range of sexually transmitted viruses are infections generally attained through uncertain sexual contact and can lead to serious health complications or may result in death if not diagnosed earlier. In this work, a Biosensor based on two-dimensional (2D) photonic crystal structures is proposed and integrated with Machine learning performance metrics. The 2D photonic crystal-based optical sensor platform produces a simulated signature outcome. The output spectral behavior varies according to the type of virus deducted. To calculate the accuracy, the signature data are trained and tested using the Machine Learning (ML) algorithm, k-Nearest Neighbors (kNN). The Modified Mach–Zehnder Interferometer (MMZI) structure shows the novelty of the work by attaining high sensitivity and quality factors of 998 nm and 3988 respectively compared to the existing work. The classification report is generated and shows a high accuracy of 97.12%. In the end, the Graphical User Interface (GUI) has been wrapped up for better readability of the results obtained.