Amazon Comprehend

Use cases:

• When you need to extract insights from an article or any type of text document
• Legal document analysis, Social media analysis
• Automatically categorize support tickets, product reviews, marketing data
• Comprehend Medical - linking medical data to medical terms and codes in standard medical ontologies
• Any deep text analysis of textual data

Features:

• Works on any text file -  PDF, Word documents, txt files

Types of analysis

• • PII - sensitive personal user data
  • Entities - main entities and concepts in the text and their position
  • key phrase - is a string containing a noun phrase that describes a particular thing
  • Dominant language - e.g. is it Japanese or English?
  • Sentiment - is it positive, negative?
  • Analyze syntax - for each word whether it's a verb or noun, etc.

• Custom training of concepts based on your business
  • If normal concepts does not fit you, you can train it "automl" style
• Can write the analysis to S3 or answer through the API

Examples

• "Daniel went to the store today"
  • Entities ->
    • "today" : date
    • "Daniel" : person

Supported languages:

German, English, Spanish, Italian, Portuguese, French, Japanese, Korean, Arabic, Chinese (simplified), Chinese (traditional)