Fake News Detection using ML & CNN

Mar 1, 2025 · 1 min read

A hybrid fake news detection system combining traditional machine learning with deep learning techniques. The project focuses on identifying misinformation by using both TF-IDF + Naive Bayes pipelines and Convolutional Neural Networks for text classification.

Key Features:

  • 🧹 Debiased the dataset to reduce skew and ensure fair model performance
  • 🔍 Built a classic ML pipeline using TfidfVectorizer and MultinomialNB for efficient fake news classification
  • 🧠 Implemented a CNN-based NLP model for deeper context understanding and semantic pattern learning
  • 🔄 Compared performance between traditional and deep learning methods to ensure robustness
  • 📊 Visualized metrics like accuracy, precision, recall, and confusion matrix for model evaluation

Technologies Used:

  • Scikit-learn (for pipeline, TF-IDF, Naive Bayes)
  • TensorFlow / PyTorch (for CNN)
  • Pandas, NumPy, Matplotlib, Seaborn (for data preprocessing & visualization)