Analyze text to extract meta-data from content such as concepts, entities, keywords, categories, sentiment, emotion, relations, semantic roles, using natural language understanding. With custom annotation models developed using Watson Knowledge Studio, identify industry/domain specific entities and relations in unstructured text.
Analyze up to 1000 items per day.
Use against any of the features.
Stores one free custom model.
1 - 250,000 items per month
$0.003 per item
250,001 - 5,000,000 items per month
$0.001 per item
5,000,000 + items per month
$0.0002 per itemCustom model price per month
$800 per modelThere is no daily limit.
Question: What are features?
Answer: Features include Categories, Concepts, Emotion, Entities, Keywords, Metadata, Relations, Semantic Roles, and Sentiment. We charge each feature in the call as a separate item.
Question: What else is included in an Item?
Answer: One item is one feature with up to 10,000 characters with a maximum file size of 50kb. For example, if you specify the service to give you Concepts and Entities on text that the service cleans to 8,000 characters, then that costs two items - for the two features. However, if the cleaned text has 11,000 characters, then this call costs you four items, because the service splits calls with more than 10,000 characters into multiple items. The service only accepts the first 50kb of a file and discards the rest.
Question: How do I create a custom model?
Answer: You can create a custom model through a subscription to Watson Knowledge Studio, and then deploy it to your instance of Natural Language Understanding.
Bluemix is IBM’s cloud platform where you can access the Watson services.
Follow the getting started guide to use Watson in five steps.
Ready to get down to the details? Full documentation detailing how to get started using this service in Bluemix is available for each Watson service.
Localized versions of Watson services (Natural Language Classifier, Retrieve and Rank, Speech to Text, Text to Speech) are available in the following places.
Japanese - SoftBank