Bust --> Breaking (down) User's Suspicious Texts

Inspiration

This year there were almost (500,000) suspicious activity reports that were received and processed by the National crime agency period (80.21% of which came from banks); there has been a 52.72% increase in the amount of suspicious activity responses lately. For this hackathon I wanted to create a webapp that aids organizations by accurately stopping this problem and finding out whose done it.

What it does

Uses progressive NLP models and sentiment analysis to detect suspicious behavior and activity through entries, tweets, texts, etc.

How we built it

Used several api's, machine learning concepts like sentiment analysis, OCR, cockrach DB

Accomplishments that we're proud of

First time using a sentiment analyzer Achieved 99.98% confidence Was able to detect lots of emotions and scale it from -1,1 Used OCR with a 100% confidence rate .. compared it with google cloud vision api to make sure accuracy is right