📰 Fake News Detector using BERT

This beginner-friendly AI mini project shows how to build a fake news detection system using the BERT model with Hugging Face Transformers.

🧰 Prerequisites

• Python 3.x
• Transformers, Scikit-learn, Pandas, NumPy
• Basic NLP knowledge

📦 Step 1: Install Required Libraries

pip install transformers scikit-learn pandas torch

🧠 Step 2: Load Dataset and Preprocess

import pandas as pd
from sklearn.model_selection import train_test_split

df = pd.read_csv('fake_or_real_news.csv')  # Dataset with 'text' and 'label' columns
train_texts, val_texts, train_labels, val_labels = train_test_split(
    df['text'], df['label'], test_size=0.2)

🧪 Step 3: Tokenize with BERT

from transformers import BertTokenizer

tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')
train_encodings = tokenizer(train_texts.tolist(), truncation=True, padding=True, max_length=512)
val_encodings = tokenizer(val_texts.tolist(), truncation=True, padding=True, max_length=512)

🎯 Step 4: Define Dataset Class

import torch

class NewsDataset(torch.utils.data.Dataset):
    def __init__(self, encodings, labels):
        self.encodings = encodings
        self.labels = labels
    def __getitem__(self, idx):
        return {key: torch.tensor(val[idx]) for key, val in self.encodings.items()} | {'labels': torch.tensor(self.labels[idx])}
    def __len__(self):
        return len(self.labels)

🤖 Step 5: Train the Model

from transformers import BertForSequenceClassification, Trainer, TrainingArguments

model = BertForSequenceClassification.from_pretrained('bert-base-uncased')
training_args = TrainingArguments(
    output_dir='./results',
    num_train_epochs=3,
    per_device_train_batch_size=16,
    per_device_eval_batch_size=64,
    evaluation_strategy='epoch',
    logging_dir='./logs',
)

trainer = Trainer(
    model=model,
    args=training_args,
    train_dataset=NewsDataset(train_encodings, train_labels.tolist()),
    eval_dataset=NewsDataset(val_encodings, val_labels.tolist()),
)

trainer.train()

🎓 Student Challenges

• Test with your own news headlines or stories
• Visualize confusion matrix
• Deploy using Flask/Streamlit
• Evaluate model accuracy

📘 More AI Projects

• 📝 Text Summarizer with Transformers
• 📸 Animal Classifier with VGG16
• 🧠 Sentiment Analyzer with BERT
• 💬 Chatbot with Python

Explore more tutorials at DarchumsTech Blog.

Try it yourself

Python Editor