Machine Learning and Metadata Tagging

Use Case 

Develop a machine based method to identify documents of security Class from a large  collection of documents with high accuracy.
The method should be robust and allow development of an automated service for security  application without the need for human supervision/effort.
 

Classification Result and summary 

Big Data classification of sample documents completed
Concept model imported into MDS Data Model
Real-time Security classifier service implemented for Classification
Classifier test result: 
~ 90% accuracy (based on a test of 1500 documents with 15 Classes)