News nowadays are fragmented and live in the moment - relevant for a day at most and then replaced by new stories. This makes it hard to keep up with so many news stories unfolding at the same time.
What it does
Our solution analyzes a big number of news articles and offers a new way to browse the news based on topics. It automatically extracts keywords from articles using TF-IDF and thus offers a unique perspective on the big picture of news stories. All of this is combined into a beautiful and intuitive UI.
How we built it
For the representation of the different generated news topics and their relations, we employed a heavily modified version of the particles.js library. This allows for a playful experience while discovering new and interesting topics. This interface offers a dynamic flow through topics leaving and new ones entering.
We developed a custom page ranking algorithm, which allows smart searching and aggregation of news articles. The TF-IDF statistical algorithm is used to extract topics from news articles.
Challenges we ran into
Due to time limitations, we had to cut back on the model development and thus couldn’t finish the topic relationship prediction.
Accomplishments that we're proud of
We’re happy with the UI concept we managed to implement. It provides a solid basis for displaying aggregated news sorted by topics. The page ranking algorithm in combination with TF-IDF statistical analysis shows promising results for AI-driven news aggregation.
What we learned
We gained a new perspective on news consumption and exploration of news topics and gained further experience with utilizing tf-idf analysis and developing page rank algorithms.
What's next for NewsRetrieval
Proper implementation of cross topic navigation.