Fake News Detection Using Scikit-Learn

This project aims to classify news articles as real or fake using a machine learning algorithm. The algorithm used is a Passive Aggressive Classifier, which is a type of online learning algorithm that is used for binary classification.

Installation

The following packages are required to run this project:

• newsapi-python
• pandas
• scikit-learn

To install these packages, run the following command:

pip install newsapi-python pandas scikit-learn

Data

The training data used to train the classifier is taken from Kaggle.

Step 1: Import Necessary Modules

The following modules are required to interact with the News API and create the model:

• NewsApiClient from newsapi
• random

Step 2: Get News Data from the News API

To interact with the News API, an API key is required. The API key can be obtained by registering at https://newsapi.org/. After obtaining the API key, follow the steps below:

1. Call the NewsApiClient() method and pass the API key to this method.
2. Create a method to get the news data from the API using the get_everything() method.
3. Pass the required parameters to the get_everything() method:
   • sources
   • domains
   • from_param
   • to
   • language
   • sort_by
   • page
4. After getting the results from the API, pass the results to an array and return that array.

Step 3: Get News Sources

The News API has over 3000 authenticated news sources. To get news from these sources, follow the steps below:

1. Get all the sources from the News API using the get_sources() method.
2. Add the ID of each source to a list.
3. Truncate the list to a size of 10, and get news from those sources using the del keyword.

Step 4: Create a DataFrame Using News List

1. Use a loop to iterate through all the sources from the sourceList and use the getNews() method to get news from the sources.
2. Add all the returned news to a list.
3. Use the from_records() method from pandas.DataFrame to create a new DataFrame using the list.
4. Add new column headings to the DataFrame using the dataframe.columns attribute.

Step 5: Create a DataFrame of Fake and Real News

1. Load the data from a .csv file available in the same directory with the name of the news.csv file using the read_csv() method from pandas.
2. Add the column headings to the DataFrame.
3. Use the concat() method from pandas to concat both DataFrames.

Step 6: Train the Model

To create the training model, the following modules are required:

• train_test_split from sklearn.model_selection
• CountVectorizer from sklearn.feature_extraction.text
• PassiveAggressiveClassifier from sklearn.linear_model
• accuracy_score from sklearn.metrics

Follow the steps below to train the model:

1. Split the training and testing data from the DataFrame using the train_test_split() method.
2. Use 70% of the data for training and 30% for testing.
3. Pass the combination of title, text, and news labels to the *arrays parameter of the train_test_split() method.
4. Use CountVectorizer to create a matrix of token count from the text document.
5. Create a PassiveAggressiveClassifier model to classify real news from fake news.
6. Test the model using the test data and calculate the model's accuracy using the accuracy_score method.

By following these steps, we can create a machine learning model to detect fake news from real news.

Credits

This project was created by OpenAI and is based on the tutorial available on DataCamp.