unable to detect health categories such as Examination, Treatment etc using Azure Text Analytics

Question

Hi Team,

I am Purnima from Citiustech Healthcare Technology Pvt Ltd and I have few queries related to Microsoft Azure Text Analytics.

I am working on one of the POC on NLP to get the health entities such as Diagnosis and Procedure from text using Azure text analytics. I have used Postman to send request. The NER feature for text analytics returns the general entity categories such as Person, PersonType, Location etc. but it is unable to detect health categories such as Examination, Treatment etc.

1) I have used below Request Endpoint in the Postman

https://<your-custom-subdomain>.cognitiveservices.azure.com/text/analytics/v3.1-preview.3/entities/recognition/pii?domain=phi

So it is identifying the Health related text such as ‘Blood Test’ into Category ‘Skill’ and not in Health Category ‘Examination’. Also how to detect only specific single Category from input text?  

2) Also If I paste 2 MB JSON data in the input it gives me below error

I request you to provide some solution on how the API can detect health related entities correctly.

Answer 1

@PurnimaJaiswal-8418 Thanks for your interest in text analytics for health. Based on the information above i think you are using the API for extracting the PII data.
Text analytics for health is a latest offering from the service which is still in preview and not recommended for production use. This page details very important information regarding its usage and applications. We would highly recommend to read through the details of this feature.

This feature since in preview needs special access to be granted for your subscription which is governed by our gating process. You can submit a request from this page and then proceed further is access is granted.

Regarding the error seen with payload limitations there is a limitation on the characters, document and requests which are also documented in the limits page. The recommendation is to use shorter requests or break the text into smaller chunks of text before sending them to the API.