🤖 AI vs Human Text Classification (NLP Project)
📌 Project Overview
This project focuses on building a Natural Language Processing (NLP) binary classification model to detect whether a given text is AI-generated or human-written.
The goal is to analyze text authenticity using machine learning techniques and evaluate model performance on unseen data.
📊 Dataset
The dataset is publicly available on Kaggle and contains labeled text samples classified as:
AI-generated text Human-written text
⚙️ Project Workflow
- Data Preprocessing Text cleaning (removing special characters, numbers, and noise) Lowercasing and normalization Preparing data for feature extraction
- Feature Extraction Bag-of-Words (CountVectorizer) representation
- Model Building
Two machine learning models were used:
Naive Bayes Classifier Logistic Regression 4. Model Evaluation Accuracy score Performance comparison on test data Prediction on unseen text samples
🧠 Technologies Used Python Pandas Scikit-learn Numpy NLP (Natural Language Processing)
🚀 Results
The models were trained and tested on unseen data, demonstrating effective classification of AI-generated vs human-written text. Logistic Regression showed better performance compared to Naive Bayes.
📁 Project Structure
├── notebook.ipynb
├── dataset.csv
├── README.md
🎯 Key Learnings Text preprocessing techniques in NLP Bag-of-Words feature extraction Binary classification models Model evaluation and comparison
📌 Author
Merna Ayman
Artificial Intelligence Graduate