From Genes to Tokens: a GWAS-inspired Approach for Interpretable Stylometric Analysis

The authors present a new way to figure out which words (called 'gene' tokens) are strongly linked to particular writers (the 'phenotypes'). They use a math technique called logistic regression, which helps see if certain words are tied to specific authors, while making sure not to get tricked by random chance. They tested this method on texts in English, German, and Russian and found words that clearly stand out as unique to individual authors. This helps better understand how an author's style can be identified through their word choices.