Binary Human-Written vs. Machine-Generated Monolingual Text Classification

Authors: Chaudhari Gauri, Kolhe Anushree

Term: Fall 2023: Indiana University

Project Repository: GitHub IUB

SemEval Task 2024

• Implemented Ensemble Classifier Approach with Stochastic Gradient Descent and XGBoost to achieve an impressive 70% accuracy in distinguishing human-written from machine-generated text.
• Integrated linguistic features with TF-IDF using a Hybrid Feature Approach for nuanced text classification, incorporating metrics like the Flesch Reading Ease score to enhance accuracy and relevance.
• Contributed to a research initiative by comparing the effectiveness of Ensemble Classifier and Hybrid Feature approaches in automated text classification.