Amazon Comprehend
Notes based on Adrian Cantrill's course from cantrill.io and Stéphane Maarek's Udemy course.
View On GitHub
Amazon Comprehend
It is a natural language processing (NLP) system available in AWS
We input a document and Comprehend will develop insights by recognizing the entities, key phrases, language, sentiments and other common elements of the document
How the system works:
Input = Document
Output = Entities, phrases, language, PII, sentiments
Comprehend is ML services and it is based on pre-trained and custom models
Product is capable of doing real-time analysis for small workloads and asynchronous analysis for large workloads in form of jobs
The text analysis type can be:
Built-in: using AWS built-in models
Custom: using custom models from endpoints we have created