DEV Community

Nimmala NAGA SANTHOSH BABA
Nimmala NAGA SANTHOSH BABA

Posted on

Create a custom entity recognition model using Amazon Comprehend

what is AWS Comprehend?

Amazon Comprehend is a natural language processing (NLP) service provided by AWS that makes it easy to discover insights and relationships in text. It can extract key phrases, entities, sentiment, and other useful information from text. Amazon Comprehend uses machine learning to analyze text and provide insights without requiring any prior machine learning experience.

Some key features of Amazon Comprehend include:

  • Entity recognition: Identify and extract key entities like people, places, organizations, dates, quantities, and more from text.
    Key phrase extraction: Identify the main themes in text documents.

  • Sentiment analysis: Determine the sentiment expressed in text, such as positive, negative, or neutral.
    Language detection: Automatically detect the dominant language in text.

  • Custom models: Build your own custom entity recognition or text classification models using your own data.

Amazon Comprehend is a powerful NLP service that can be integrated into various applications to extract valuable insights from text data. It's a great tool for tasks like content analysis, customer support, and information extraction.

What is the business case ?

1.Create a custom entity recognition model using Amazon Comprehend
You will upload training data to an Amazon S3 bucket, then use Amazon Comprehend to train a custom entity recognition model based on that data.

2.**Conduct a real-time analysis with a custom entity recognition endpoint. Once the model is trained, you will create an endpoint to perform real-time analysis on documents, identifying the custom entities you defined during the training process.

Architecture :-

Image description

need to create s3 bucket and hold 2 files document.txt and annotations.txt like below

Image description

now launch Amazon Comprehend and create a new model like below.

Image description

Image description

Image description

Image description

Image description

once you submit this, it takes a few minutes to train the data

Image description

you can see progress here

Image description

after a few minutes, the model will be trained like below.

Image description

create an endpoint for this model to test

Image description

Image description

Image description

In the navigation pane at the left of the page, choose Real-time analysis.

Image description

pass this data like below .

Image description

finally you learned about AWS comprehend , with this practice lab.

Top comments (0)