How we did it
- Identify which external service to leverage to obtain latest news stories from everywhere
- Ignore irrelevant results
- Remove exact or near duplicate news stories
- Determine relevance and sentiment of each news story
- Handle large volumes
- Ensure quick results
- Data security and privacy
The solution was setup to ensure scalability with the expectation of a large volume of query. The entire setup was made in the cloud fully controlled by the client.
- Distributed, scalable architecture
- Core natural language understanding engine
- Clustering of news stories based on their content
- Analysis of trends in the news to help recognize, classify and disambiguate input query terms as needed
- Client cloud control
Each news result is ranked by importance and with an associated overall sentiment score. Main topics in each of the news results together and corresponding sentiment score is also returned. The real-time API returned results almost immediately creating an elevated user experience and increasing the apps stickiness.
The success of using MindWave AI led to deeper engagement on a number of fronts with the client.