Week 12

21 April :  Style Experiments (continued)

 

Subject:  This week we will look more deeply into ways of carrying out computational experiments with style designed to help you carry out your own small experiment that is the subject of blog 4 (NB: extended due date).

Format:  Class on Tuesday will be a synchronous meeting walking us through the functionality of the Stylo GUI.  There will be no quick writing this week. There will be no synchronous class on Thursday.  All work the rest of the week is asynchronous.

Notebook:  The notebook for today’s class is the one we used last week, and it is located in Drive.

 

***

Synchronous class schedule:

We will work with 3 corpora that are in the class folder (AustenBrontes, LittleCousin, Doyle)

Intro / downloading corpora
Opening last week’s notebook
Beginning a new notebook

Working first with commands in RStudio and then with settings in the Stylo Graphic User Interface (GUI), beginning at step #4

Cluster Analysis (Little Cousin) – lines 119 on
Bootstrap Consensus Tree (AustenBrontes) lines 129 on

Who was Arthur Conan Doyle?

Stylo options:

1   FEATURES = numbers of words, chunks
2   STATISTICS = kinds of analysis and different distance measures
3   OUTPUT = different modes of visualization
4   Kinds of files that are generated in the corpus folder

 

***

 

Ideas for blog 4:

-Find an author in PG that is disputed and try to pair them with other authors they are believed to be.
-Find an author in PG who has written many books.  How well does a bootstrap consensus tree distinguish between that author’s different genres, time periods.
-Try combining sampling with some very long novels along with Principle Component Analysis to see if style varies across different parts of books.
-Try the corpus of as much of your own writing as you can put into txt format.  -How does stylometric analysis classify your own different genres of writing?  Take another person’s writing and slip it several 1000 words in to some of your essays. Can stylo detect this? Try chunking it.
-Try a corpus of some papers written by different friends from three countries.  Are you able to determine a national signal in their writing?

 

VOH:  Wednesday and Thursday this week by sign up

NB: VOH change to evening on Wed and Thurs with the beginning of Ramadan.