Automated Recommendation Feed for Hacker News
This project automates the recommendation feed from Hacker News using Naive Bayes from scikit-learn. It parses the Hacker News website to populate the local database. The project consists of two parts: a basic feed where users can like and dislike news articles, and a personalized feed built upon the user's preferences. The recommendation system uses NBC (Naive Bayes Classifier) to calculate the likelihood of a particular article being interesting to the user.
To get started with the Personalized HackerNews Feed, make sure you have Python and the required dependencies installed. Clone the repository to your local machine and follow the installation instructions below.
Use the following commands to set up the project:
# Clone the repository
git clone https://github.com/pol3et/automated-feed.git
# Navigate to the project directory
cd automated-feed
# Install the required dependencies
pip install -r requirements.txtTo use the project, follow these steps:
- Parse HackerNews to populate the local database:
python populate_db.py- Run the local server:
python hackernews.py- Open your web browser and visit localhost:8080/news to like or dislike articles and train the recommendation model. Then, navigate to localhost:8080/feed to see personalized suggestions.
- Automatically scrapes and renders website into a database.
- Utilizes Naive Bayes from scikit-learn.
- Provides a simple and effective base to build upon.
This project was completed as part of the tasks from Pybook by Dementiy.
