Google News Recommendation System working is what inspired us to do this project.
- Recommends the Top-10 headlines based on the user's interest or choice of reading
- We give the most relevant ones making them suitable for users to select.
- Python
- Jupyter Notebook
- Machine Learning
- Streamlit
- We tried to work on abstract summarization but due to a few errors in LSTM Model, we were unable to proceed.
- Also there were difficulties leveraging the IBM Linux One Platform. But consistent support from mentors helped us overcome these difficulties.
- We were able to get the most relevant search results for any user-entered article headline.
- We are able to recommend the headlines based on the articles.
- Create a front-end so that it is not only notebook accessible but could be used by any normal users.
- Display the articles with their short summaries which eases the choosing time for the user.
- Different Python Modules are used to do NLP-based processing
- Learned to implement our modules in IBM Cloud-based platform
- Different steps involved in NLP-based projects
- Build a SaaS product and make it open-source for everyone to access and stay updated.
- Working to integrate Web Scraping to support different other news categories.
Note: Main file that contains the final code is bbc_prod.ipynb
