Comprehension of the Shakespeare authorship question through deep impostors approach.

This article presents a novel computational methodology, termed the "Deep Impostor" approach, to investigate the Shakespeare authorship question by analyzing stylistic patterns in texts attributed to William Shakespeare. The method involves segmenting texts into equal word batches, training deep neural networks—specifically Convolutional Neural Networks (CNN) and pre-trained BERT transformers—to distinguish between pairs of known impostor texts, and then transforming target texts into numerical signals based on these classifications. Using Dynamic Time Warping distance and Isolation Forest anomaly detection, the approach clusters Shakespearean works into groups, identifying a subset of fifteen texts as stylistically inconsistent with Shakespeare's authentic style, suggesting possible alternative authorship or co-authorship. The study's CNN-based results align with some scholarly debates on disputed works, while the BERT-based model showed less definitive outcomes, attributed to differences in training procedures.