Machine-Learning Classification of Closed and Open Radiating Wires from Near Magnetic or Electric Field Scan Images
This work addresses the need for automated identification of radiating coupling sources in electromagnetic applications, but it is incremental as it applies existing methods to a specific domain.
The researchers tackled the problem of automatically classifying the shape of radiating wires from near-field scan images by applying machine learning classifiers like SVM, k-NN, and Gaussian processes, achieving performance validated through leave-one-out cross-validation.
Sets of intelligent classifiers are applied to the near-field scan-data in order to automatically classify the shape of radiating wirings. The support vector machine, k-nearest neighbors algorithm, and Gaussian process classifications are trained using the near-field radiation pattern of diverse radiating wire configurations. Leave-one-out cross-validation is used for estimating the performance of the predictive models. The output of this research is a software package well-suited to be retrained based on any measured near-field databank to automate the identification of magnetic-type or electric-type of the radiating coupling sources.