Text Categorization - An Evaluation of Textual Features for Automated Genre Classification

Thesis Type Master
Thesis Status
Currently running
Student Sonja Wirtenberger
Start
Thesis Supervisor
Contact
Research Field

The goal of this thesis is to evaluate how well text features and their combinations perform at text categorization tasks, with focus on genre classification. A tool for automated testing should be developed, which imports datasets, extracts the desired text features and executes the tests using different classification algorithms. The results should be shown on a webpage, where users have the possibility to categorize their own texts based on a trained classifier.