📰 Fake News Detection Using Machine Learning

🔍 Overview

This project addresses the rising challenge of online misinformation by building machine learning models to identify fake news articles. We used NLP techniques and multiple classifiers to predict whether a news article is real or fake based solely on its content.

🧭 Approach

We collected and cleaned a labeled dataset of news headlines and bodies. The data was vectorized using TF-IDF and passed through various supervised learning models to evaluate performance in classifying articles.

⚙️ Methodologies

• Text Cleaning: Removed punctuation, stopwords, and special characters using regex
• TF-IDF Vectorization: Converted text to numerical vectors
• Modeling: Trained Logistic Regression, Naive Bayes, Decision Tree, SVM, Random Forest, and Gradient Boosting
• Evaluation: Accuracy, confusion matrix, precision, recall, F1-score

🧰 Technologies

• Language: Python
• Libraries: Scikit-learn, Pandas, NumPy, Matplotlib, Seaborn
• NLP Techniques: TF-IDF, text preprocessing
• Models: Logistic Regression, Naive Bayes, SVM, Random Forest, Gradient Boosting

💡 Key Learnings

• Gained hands-on experience applying NLP techniques to real-world data
• Learned how to evaluate model performance for high-stakes classification problems
• Understood how different algorithms perform on imbalanced text datasets

📈 Results

Logistic Regression and Gradient Boosting yielded the highest accuracy in classifying fake news, supported by precision-recall analysis. The model demonstrated robust performance across various test cases, making it a valuable tool for identifying misinformation.