New – Train Custom Document Classifiers with Amazon Comprehend

Amazon Comprehend gives you the power to process natural-language text at scale (read my introductory post, Amazon Comprehend – Continuously Trained Natural Language Processing, to learn more). After launching late 2017 with support for English and Spanish, we have added customer-driven features including Asynchronous Batch Operations, Syntax Analysis, support for additional languages (French, German, Italian, and Portuguese), and availability in more regions.
Using automatic machine learning (AutoML), Comprehend lets you create custom Natural Language Processing (NLP) models using data that you already have, without the need to learn the ins and outs of ML. Based on your data set and use case, it automatically selects the right algorithm, tuning parameter, builds, and tests the resulting model.
If you already have a collection of tagged documents—support tickets, call center conversations (via Amazon Transcribe, forum posts, and so forth)— you can use them as a starting point. In this context, tagged simply


Original URL: http://feedproxy.google.com/~r/AmazonWebServicesBlog/~3/gNJLej3dlEA/

Original article

Comments are closed.

Proudly powered by WordPress | Theme: Baskerville 2 by Anders Noren.

Up ↑

%d bloggers like this: