Bayliss Wagner's Projects
Author Profiling ToolKit.Implementation of my author profiling tools as an integrated tool which uses spacy and scikit learn to analyze text train and classify.
This was a group final project for our CS65 Natural Language Processing course. We built a naive Bayes baseline and Bayesian classifier in order to see if we could train the latter to identify author gender. We used two different data sets: a set of NYT Opinions articles, and a set of miscellaneous blog posts. Each had an equal number of men and women. In addition, we built a doc2vec system, which created vectors for each document in the corpora and determined which were most likely male- or female-authored. We presented on our project and documented our results in a paper conforming to Association of Computational Linguistics conference standards; you can see our paper & slides in the 'paper' directory.
This program calculates combinations and permutations using C++. I made it while I was teaching probability this summer because there was no wifi in my classroom and I didn't always have time to calculate the answers to problems before classtime!
First website prototype for 2019 DJNF digital media residency program on development boom in Downtown Phoenix.
personal website
This Python file, created for a Digital Humanities assignment (woohoo), takes single novels or corpora of novels as input. The user can choose novels from these files and compare them. The program displays "most frequent words" in each novel and "most distinctive words" in each through comparison. The program also displays the "context" (surrounding words) of any word in the novel. (Ex: input "love" will display every use of the word "love" in the novel). I experimented with removing different stopwords when we returned to this program for my English class.
Final project for CS63 Artificial Intelligence, with partner. We trained a neural network to classify recipes by type of cuisine (Indian, French, etc.) based on their ingredients.