Stylometry is analysis of textual data to find hidden patterns. This paper provides a comprehensive survey on this topic.

The paper contains three main parts. In the first part, the authors consider five stylometry problems:

• Authorship attribution: finding the author of a text in a closed/predefined (or open) set of authors;
• Authorship verification: determining whether given documents were written by the same author;
• Authorship profiling: obtaining author demographics such as gender;
• Stylochronometry: assigning time to text; and
• Adversarial stylometry: altering style for evasion of authorship.

In the second part of the paper, they provide performance analysis for large-scale authorship analysis using 14 algorithms. They show that with several authors, this important task is challenging and invite new researchers to the field. In the third part, they define open problems.

The authors describe several open-source and commercial systems. The reference list is rich, with some omissions, and covers studies in different languages. It contains tables for feature categories, publicly available resources in terms of datasets and software, and so on. The tables make the material more accessible.

I like the emphasis of the large-scale authorship analysis that involves more than 1,000 authors and limited data. I see this aspect of stylometry as important in today’s world, with several possible dangers on the web and social media. The paper will be useful for practitioners and researchers. I enjoyed it.