Amazon Comprehend
Amazon Comprehend is a natural language processing (NLP) service that uses machine learning to uncover information in unstructured data. The service can identify critical elements in data, including references to language, people, and places. It provides entity recognition, sentiment analysis, key phrase extraction, language detection, topic modeling, PII detection, and custom classification capabilities for text documents at scale.
APIs
Amazon Comprehend API
The Amazon Comprehend API provides programmatic access to natural language processing capabilities including sentiment analysis, entity recognition, key phrase extraction, langu...
Capabilities
Amazon Comprehend NLP Analysis
Workflow capability for natural language processing analysis including entity recognition, sentiment analysis, key phrase extraction, language detection, PII detection, and cust...
Run with NaftikoFeatures
Identify named entities such as people, places, organizations, dates, and quantities in text.
Detect positive, negative, neutral, or mixed sentiment in text documents.
Extract key noun phrases and concepts from unstructured text.
Automatically detect the dominant language of input text from 100+ supported languages.
Identify and optionally redact personally identifiable information from documents.
Discover latent topics in large collections of documents using unsupervised learning.
Train custom document classifiers using your own labeled training data.
Train custom entity recognizers to identify domain-specific entities in text.
Analyze the syntax of text including parts of speech tagging and tokenization.
Continuously improve custom models with new training data using the Flywheel feature.
Use Cases
Analyze customer reviews, support tickets, and survey responses to understand sentiment and extract actionable insights.
Detect and filter inappropriate content, hate speech, or PII from user-generated content.
Extract entities, key information, and classifications from contracts, reports, and medical records.
Mine customer interactions for trends, topics, and sentiment to improve products and services.
Automatically detect and redact PII from documents to meet GDPR, HIPAA, and other compliance requirements.
Enrich search indexes with entities, key phrases, and topics to improve search relevance.
Integrations
Process documents stored in S3 buckets using async batch analysis jobs.
Integrate custom Comprehend models with SageMaker for advanced ML workflows.
Trigger real-time Comprehend analysis from Lambda functions in event-driven architectures.
Analyze streaming text data from Kinesis Data Streams and Firehose.
Enrich OpenSearch indexes with Comprehend-detected entities and topics for better search.
Use Comprehend in AWS Glue ETL jobs for large-scale document processing pipelines.