Deep Learning-based Algorithm for Detecting Counterfeit Domain Names

Abstract

To address the two major problems of machine learning: the high cost of manually extracting features and the requirement for reasonable and high relevance of feature set input, this paper will use the ability of deep learning to automatically characterize and learn data to build a lightweight counterfeit domain name deep learning grid-based detection model based only on domain name strings, and conduct comparison experiments with edit distance-based detection models and visual feature-based detection models by testing on public datasets to verify the effectiveness of this deep learning-based counterfeit domain name detection model for counterfeit domain name detection.