SkillAscend is an AI-powered online course recommender system that provides personalized course recommendations based on a user's educational background and learning goals. It leverages a hybrid recommendation approach by combining collaborative filtering (SVD) and content-based filtering (BERT & TF-IDF) to suggest the most relevant courses dynamically.
- Personalized Recommendations: Matches users with courses based on their education and career goals.
- Hybrid Recommendation Approach:
- Collaborative Filtering (SVD): Predicts user preferences based on historical interactions.
- Content-Based Filtering (BERT): Recommends courses by analyzing course descriptions and user profiles.
- Real-Time NLP Pipeline: Updates recommendations dynamically as users interact with the system.
- User-Friendly Web App: Simple and interactive UI for seamless course discovery.
- ML: TensorFlow, scikit-learn, Surprise (SVD), Sentence Transformers (BERT), TF-IDF, Pandas, NumPy
- Backend: Flask, Hugging Face Spaces
- Frontend: React.js, Tailwind CSS, HTML
SkillAscend/
│── data/ # Dataset (Coursera courses & interactions)
│── models/ # Trained models (SVD, BERT, TF-IDF)
│── backend/
│ ├── app.py # Flask backend for recommendations
│ ├── realtime_nlp.py # Real-time NLP-based recommendation
│ ├── collaborative_filtering_svd.py # SVD-based recommendation
│ ├── content_filtering_bert.py # BERT-based content filtering
│ ├── content_filtering_tfidf.py # TF-IDF-based content filtering
│── frontend/
│ ├── index.html # Web UI for recommendations
│ ├── static/ # CSS, JavaScript, images
│── README.md # Project documentation
│── requirements.txt # Python dependencies
- Clone the Repository:
git clone https://github.com/your-username/SkillAscend.git cd SkillAscend - Install Dependencies:
pip install -r requirements.txt
- Run the Backend:
python backend/app.py
- Access the Web App:
Open
http://127.0.0.1:5000/in your browser.
- User inputs their education and learning goals.
- BERT & TF-IDF process the user profile and compute similarity with course descriptions.
- SVD-based collaborative filtering predicts courses based on past user interactions.
- The top 5 courses are recommended and displayed to the user.
- Support for multiple course providers (Udemy, edX, etc.)
- User authentication & profile saving
- Improved ranking with hybrid weighting techniques
- Integration with APIs for real-time course updates