Fake News Detection Project

This project demonstrates a simple fake news detection model using a Passive Aggressive Classifier and TF-IDF vectorization.

Project Steps:

1. Load Data: The project starts by loading the news data from a CSV file into a pandas DataFrame.
   

df = pd.read_csv('/content/news.csv')
print(df.shape)
display(df.head())

# Get accurate labels from dataframe
labels=df.label
labels.head()

Below is the training set used for the model and output:

Below is a preview of the dataset used for training and testing the model:

| Unnamed: 0 | Title                                         | Text                                                        | Label |
|------------|-----------------------------------------------|-------------------------------------------------------------|-------|
| 8476       | You Can Smell Hillary’s Fear                  | Daniel Greenfield, a Shillman Journalism Fello...            | FAKE  |
| 10294      | Watch The Exact Moment Paul Ryan Committed... | Google Pinterest Digg Linkedin Reddit Stumbleu...            | FAKE  |
| 3608       | Kerry to go to Paris in gesture of sympathy   | U.S. Secretary of State John F. Kerry said Mon...            | REAL  |
| 10142      | Bernie supporters on Twitter erupt in anger...| — Kaydee King (@KaydeeKing) November 9, 2016 T...            | FAKE  |
| 875        | The Battle of New York: Why This Primary...   | It's primary day in New York and front-runners...            | REAL  |

2. Split Data: The data is split into training and testing sets to prepare for model training and evaluation.
   

# Split into training and testing sets
x_train,x_test,y_train,y_test=train_test_split(df['text'], labels, test_size=0.2, random_state=7)

3. Vectorize Text: The text data is transformed into numerical features using TF-IDF (Term Frequency-Inverse Document Frequency) vectorization. This process converts the text into a matrix of token counts, weighted by their frequency.
   

# Initialize a TfidfVectorizer
tfidf_vectorizer=TfidfVectorizer(stop_words='english', max_df=0.7)

# Train and transform the training set, and transform test set
tfidf_train=tfidf_vectorizer.fit_transform(x_train)
tfidf_test=tfidf_vectorizer.transform(x_test)

4. Train Model: A Passive Aggressive Classifier is initialized and trained on the vectorized training data. This is a type of online learning algorithm that is suitable for large datasets.
   

# Initialize the PAC
pac = PassiveAggressiveClassifier(max_iter=50)
pac.fit(tfidf_train, y_train)

5. Evaluate Model: The trained model is used to make predictions on the test set, and the accuracy of the model is calculated. A confusion matrix is also generated to understand the performance in terms of true positives, true negatives, false positives, and false negatives.
   

# Make predictions on the test set and calculate accuracy
y_pred = pac.predict(tfidf_test)
score = accuracy_score(y_test, y_pred)
print(f"Accuracy: {round(score*100,2)}%")

Output:

Accuracy: 92.58%

Code Highlights:

• Loading data using pandas.read_csv().
• Splitting data into training and testing sets using sklearn.model_selection.train_test_split().
• Initializing and fitting a TfidfVectorizer from sklearn.feature_extraction.text.
• Initializing and training a PassiveAggressiveClassifier from sklearn.linear_model.
• Evaluating the model using sklearn.metrics.accuracy_score() and sklearn.metrics.confusion_matrix().

Results:

The model achieved an accuracy of 92.2% on the test set.

	Predicted FAKE	Predicted REAL
Actual FAKE	586 (TP)	43 (FN)
Actual REAL	51 (FP)	587 (TN)