Fake News Prediction Model

In this model, we will give a labeled data set. It consists of several thousands of news articles.

Labels are: -

• Fake News
• Real News

In this, we will use a Logistic Regression Model. Logistic Regression Model is?a data analysis technique that uses mathematics to find the relationships between two data factors. It then uses this relationship to predict the value of one of those factors based on the other. The prediction usually has a finite number of outcomes, like yes or no.

It is a binary classification problem.

There will be only 2 outputs whether fake news or real news.

WORKFLOW

• News Data
• Data Preprocessing
• Train Test Split
• Logistic Regression Model
• New data to Trained Logistic Model

Dataset Link: - https://www.kaggle.com/competitions/fake-news/data?select=submit.csv

GitHub Link: - https://github.com/dvvansia1805/Fake-News-Prediction-Model/tree/main

import numpy as np
import pandas as pd
import re
from nltk.corpus import stopwords
from nltk.stem.porter import PorterStemmer
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score

import nltk
nltk.download('stopwords')

# Print the list of stopwords
print(stopwords.words('english'))

# Load the dataset using a properly formatted path
news_dataset = pd.read_csv(r'D:\College Work\AI\Fake News Prediction\train.csv')

# Print the shape of the dataset
print("Shape of the dataset:", news_dataset.shape)

# Display the first few rows of the dataset
print("First few rows of the dataset:")
print(news_dataset.head())

# Check for missing values
print("The number of missing values in the dataset:")
print(news_dataset.isnull().sum())

# Fill missing values in 'author' and 'title' columns with empty strings
news_dataset['author'].fillna('', inplace=True)
news_dataset['title'].fillna('', inplace=True)

# Combine the author and title columns to create a single text column
news_dataset['content'] = news_dataset['author'] + ' ' + news_dataset['title']

# Display the first few rows of the new dataset with the content column
print("First few rows of the dataset with the new content column:")
print(news_dataset[['content']].head())

# Data preprocessing
# Initialize the PorterStemmer
port_stem = PorterStemmer()

# Function to preprocess the text
def preprocess_text(text):
    text = re.sub('[^a-zA-Z]', ' ', text)  # Removing all the special characters and numbers
    text = text.lower()  # Converting to lowercase
    text = text.split()  # Splitting into words
    text = [port_stem.stem(word) for word in text if not word in stopwords.words('english')]  # Stemming and removing stopwords
    text = ' '.join(text)  # Joining the words back to form a single string
    return text

# Apply the preprocess_text function to the content column
news_dataset['content'] = news_dataset['content'].apply(preprocess_text)

# Display the first few rows of the dataset after preprocessing
print("First few rows of the dataset after preprocessing:")
print(news_dataset[['content']].head())

# Separating the data and label
X = news_dataset['content'].values
y = news_dataset['label'].values  # Assuming there's a 'label' column

# Convert the textual data to numerical data
vectorizer = TfidfVectorizer()
X = vectorizer.fit_transform(X)

# Splitting the dataset into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, stratify=y, random_state=42)

# Initialize the Logistic Regression model
model = LogisticRegression()

# Train the model
model.fit(X_train, y_train)

# Make predictions on the test set
y_pred = model.predict(X_test)

# Calculate the accuracy score of test data
accuracy = accuracy_score(y_test, y_pred)
print(f"Accuracy: {accuracy * 100:.2f}%")

X_new = X_test[0]
def predict_news(model, X_new):
    # Make predictions
    prediction = model.predict(X_new)
    
    # Print the prediction
    print("Prediction:", prediction[0])
    
    # Interpret the prediction
    if prediction[0] == 0:
        print('The news is Real')
    else:
        print('The news is Fake')

# Example usage:
# Assuming X_new is a single new data point
X_new = vectorizer.transform(["New news article to predict"])
predict_news(model, X_new)