KJR was engaged to design and implement a solution for classifying large volumes of scientific documentation to improve document visibility and availability for wider use. Utilising the AWS technology stack KJR achieved this through the development of a Natural language Processing classification model, that classifies scientific text topics.
The project approach was divided into the below project phases:
- AWS (AMAZON WEB SERVICES) BUILD AND TOOL CONSUMPTION
- MODEL CREATION, ACCURACY & TIME SAVED
- DETERMINING SUCCESS WHERE NO GROUND TRUTH EXISTS- HUMAN IN THE LOOP
Read the downloadable 3-part case study attached at the top of this page for full details.