A notebook that leverage the power of NLP and Unsupervised learning in order to extract the most asked question from a huge corpus of question.
This project focuses on using advanced natural language processing techniques to extract and cluster frequently asked questions from a company's website. By generating an initial set of 130 potential questions, representing them numerically with pretrained BERT models, calculating their semantic similarity, and applying HDBSCAN clustering, we aim to automate the creation of a comprehensive FAQ section, ultimately improving customer support and enhancing the user experience on the company's website.
This project is built with a powerful stack of tools, including :
- Python
- PyTorch
- SentenceTransformer (BERT Model)
- T-SNE for Dimensionality reduction
- HDBSCAN for Clustering
- Bokeh for Data visualization
This is the 10 frequently asked questions using the HDBSCAN clustering result :
| Cluster name | Questions |
|---|---|
| 1 | Can I choose my room preferences when I book? Can I get assistance with tour bookings or tickets? Can I arrange for a wake-up call in the morning? Can I store valuable items in a hotel safe? Can I change my room type upon arrival if available? Can I arrange for early morning coffee or tea delivery? Can I have flowers or gifts delivered to my room? Can I use a hotel phone for local calls? Can I upgrade my room after checking in? Can I book a room for an extended period such as a month? Can I request a late-night check-in or early check-out? Can I get assistance with planning local sightseeing tours? Can I book a conference room or business meeting space? Can I have a private event with personal catering? Can I receive detailed maps of local walking tours? Can I have a personal trainer for in-room workouts? |
| 2 | What's your policy on early check-in requests? What's the hotel's policy on lost and found items? What's the policy on lost or misplaced room keys? What's the procedure for early check-out? What's the policy for receiving mail or packages? What's your policy on accommodating dietary restrictions for meals? |
| 3 | Are there designated smoking areas on the property? Are there any nearby scenic walking or jogging trails? Are there any nearby scenic walking or jogging trails? Is there a designated area for picnics or outdoor dining? Is there a designated area for meditation or yoga? Is there a designated area for art or craft workshops? |
| 4 | How can I access the hotel's full menu of services and amenities? How can I access the hotel's digital concierge services? How can I access the hotel's art or sculpture gallery? How can I access the hotel's library of books or movies? How can I access the hotel's VIP or concierge lounge? |
| 5 | Is there a fitness facility or gym within the hotel premises? Is there a pool or spa on the hotel premises? Is there a library or reading room within the hotel? Is there a quiet area for reading or studying within the hotel? |
| 6 | Is there a quiet time policy in the hotel? Is there a curfew or quiet hours in the hotel? Is there a loyalty program for frequent guests? Is there a car rental service available at the hotel? |
| 7 | Is there an airport shuttle service available? Is there a courtesy shuttle to popular tourist attractions? Is there a shuttle to nearby golf courses? |
| 8 | Do you offer any special packages for events or celebrations? Do you offer assistance for planning proposals or special occasions? Do you provide special amenities for honeymooning couples? |
| 9 | Do you offer laundry or dry-cleaning services? Do you offer babysitting or childcare services? Do you provide transportation to local hospitals or medical facilities? |
| 10 | Can I request a specific floor or room location? Can I schedule wake-up calls with specific music or sounds? Can I schedule room cleanings at specific times? |
- Yassine DRISS - Final year AI and Data science student - Yassine DRISS