Anonymous Author Matcher

Analyses the latent clues in writing styles and identifies the author if the database has content written by them.

Dataset

• Posts and comments from 50 redditors are processed as per dataset_maker.ipynb and written to ./data.
• Reusable ETL pipeline for training is in redditors_comments_dataset.py.

Modeling Strategy

• Train a sequence encoder model to distinguish whether the two input sequences are written by same author by setting it up as a binary classification task.