AI vs Human Text Classification with K-Fold Cross-Validation

Version 0.1.0: AI vs Human Text Classification with K-Fold Cross-Validation

Features

1. K-Fold Cross-Validation Implementation:

   • Added support for 5-Fold Cross-Validation to validate the performance of the Logistic Regression model.
   • Evaluates metrics like Accuracy, Precision, Recall, and ROC AUC for each fold.
   • Provides mean and standard deviation for all metrics to ensure model robustness.

2. Logistic Regression Optimization:

   • Trains the model with TF-IDF vectorized text features.
   • Final model trained on the full dataset after cross-validation.

Visualizations

• Histogram of Decision Function Scores:

  • A histogram visualizing decision scores and the classification decision boundary.

• Top Features for AI vs Human Classification:

  • Bar charts showing the top 10 positive (AI-indicative) and negative (Human-indicative) features.

• Precision-Recall Curve:

  • A plot demonstrating the trade-off between precision and recall for Logistic Regression.

Code Optimizations

• Streamlined preprocessing functions to clean text efficiently without external dependencies.
• Optimized hyperparameters for Logistic Regression for better performance during cross-validation.

---

Getting Started

1. Clone the repository:

   git clone <repository-url>