• Contents:

Amazon Comprehend - Overview

• 🔧 For Natural Language Processing (NLP)
  • Real-time or async analysis
• Fully-managed and serverless service
• Uses ML to find insights and relationships in text
  • Language of the text
  • Extracts entities (key phrases, places, people, brands, events…)
  • Understands how positive or negative the text is (sentiment)
  • Analyzes text using tokenization and parts of speech
  • Automatically organizes a collection of text files by topic
• Sample use cases:
  • Analyze customer interactions (emails) to find what leads to a positive or negative experience
  • Create and group articles by topics that Comprehend will uncover

Custom Classification

• Organize documents into categories (classes) that you define
  • e.g. categorize customer emails (to then take possible actions)
• You define classes, train Comprehend with your data, and Comprehend then classifies documents
• Architecture Diagram

Custom Entity Recognition

• Named Entity = Predefined, general-purpose entity (people, places, organizations, dates…)
  • NER → Named Entity Recognition
  • NER provided by default in Comprehend
• Comprehend can recognize custom entities (policy numbers, phrases that imply a customer escalation…), but needs to be trained
  • Train model with custom data that contains those entities
• Architecture Diagram
• Screenshot