The paper introduces a computer‑assisted authorship‑attribution technique that employs character‑level n‑gram language models grounded in information‑theoretic principles, reporting state‑of‑the‑art results on Greek, English, and Chinese corpora with an 18 % accuracy gain for the Greek set. It contributes to authorship attribution and stylometric profiling by offering a language‑independent approach that avoids extensive preprocessing or feature selection.
Bears on: authorship & stylometry · concepts & methods