Рет қаралды 4,758
In this video, I describe how stylometry and PCA can be used to study a corpus of texts. While I focus on Chinese texts, these methods are broadly applicable to works in any language.
The figures shown here, and all data processing, was originally done using custom Python scripts and data used for culturalanalyti... You can find the raw data and scripts at dx.doi.org/10.7...
It will probably be evident that I've drawn significant inspiration from the Computational Stylistic Group's work (and their R package "stylo" in particular) and work done by scholars such as Karina van Dalen-Oskam, John Burrows, and Christof Schöch (and many others).
I created this visualization using Processing 3, Audacity, and iMovie (and my dad's old Snowball mic).
Thanks to all who commented on an earlier version of this video! Your feedback was extremely helpful!