Skip to content

Latest commit

 

History

History
2 lines (2 loc) · 453 Bytes

README.md

File metadata and controls

2 lines (2 loc) · 453 Bytes

This is a Cross-lingual study on Authorship Identification. I implemented the concept using Deep Learning Methods (CNN and Fully Connected Neural Network) and Machine Learning (SVM). The program inputs various authors that have written documents in multiple languages, and a test documents. The output is the predicted author for the given test document.
This program is implemented using CNN and main.ipynb is the file that I use for running tests.