The current implementation of the API provided by Owlin lacks interpretation; the articles are selected in a random fashion while the information in the article could be used very efficiently to determine a good ordering.
What it does
Recommend articles based on certain preset criteria.
How we built it
Challenges we ran into
We had never dealt with natural language before. This definitely was a challenge.
Accomplishments that we're proud of
What we learned
What's next for KPMG_case
Implement a deep learning algorithm to analyse the articles.