Text Analytics
Welcome to the Text Analytics Forum
Categories
API – Any ideas or feedback pertaining to features or enhancements to Text Analytics API.
Documentation – Any ideas or suggestions for the API Reference or Documentation.
Language Support – Submit a request to have a particular language supported.
Samples & SDK Request – Let us know if you would like to see a Code sample or SDK provided.
Attention!
We have moved our Customer Feedback & Ideas for Azure Cognitive Services portal to the Azure Feedback Forum.
-
1 vote
-
scanned documents OCR
In case of scanned documents, If you can recognize the type of document at least among the most commonly used ones e.g. CVS, Walgreen pharmacy drug labels based on pre well defined formats , It would be much better OCR returned!
1 vote -
FISMA
I consult to a government agency. We work with petabytes of data on multiple sources of data internal to the agency. We need to do an Entity recognition that can determine when mixing data sources, the data does not go to a High, Classified or Sensitive level from a lower level (Low, Moderate, unclassified).
This is similar to the PII NER looking for bank account numbers, SS#s etc. However, the review of mixed data in Key Phrases, should respond what FISMA level via confidence level the row of data is, or the column of data.
1 vote -
Improving accuracy in Named Entity Recognition
Hi
I currently work with a Fintech start-up and we run into a compliance requirements to remove personal names from a bunch of texts.
We resorted to looking into Azure ML and yes, it appears promising by delivering a c. 80% accuracy in screening out individual names.
Yet - it's not perfect. We have a database of hundreds of thousands of verified names and can contribute if needed.
It would be grateful if we will be approached by a Microsoft person and we can share this database if see fit. trust the wealth of data we have would greatly improve…
1 vote -
Add Container Image for Named Entity Recognition (NER)
There are container images for 3 services currently; one of them is key phrase extraction, but that is not nearly as powerful as NER. It would be great to have NER also come in a container for running on-prem without having to send the actual data to Azure.
2 votes -
Why input is required strictly in JSON format?
Any thoughts why this isn't simplified for the most common use cases? I have txt/docx/pdf/image file as input and all the complexities or converting to structured format should be abstracted away from user. Isn't this a very common use case that you have several unstructured or different types of files present in blob containers, let the API crawls through it and present its results. Similar to what Azure Search does? Also, why don't you unify these services, meaning the text analytics in Azure ML recognizes only 3 NER types, whereas Azure cognitive services recognizes more. Shouldn't this all be internally…
2 votes -
Francisco Franco is not the spanish "jefe del estado" anymore
If you enter "Clamor empresarial en presencia del jefe del Estado para poner fin cuanto antes a la violencia en Cataluña" in the test page, Text Analytics returns Francisco Franco!!! as "jefe del Estado".
This guy died in 1975.1 vote -
1 vote
-
Adding custom Named Entity Recognition to Text Analytics
We need the ability to add custom entities to Text Analytics NER (Named Entity Recognition) than what is supported. For example: I want to add a skill or a quality measure to the text posted by a user.
8 votes -
Support stock market tickers
Support basic level of stock market tickers. If I were to analyze, "$SPY will be up five points by end of week" the API believes it is film related. Relate "up five points" as a positive sentiment in relation to "$SPY" or any other ticker.
Thanks,
1 vote -
Identify location and travel types and subtypes for Named Entity Recognition
Locality "here" Map coordinates if have access to location
Locality "nearby" within 0.5 miles
District "London" - Geofence for London
Affiliated address"Mum's house"
PlaceOfWork" - your workplace
Address "Barclay's Head Office"
Bus stop
Motorway Junction "M25 J4"
Road Vector "A256"
Road Junction "corner of 6th and 12th street"
Bridge "London Bridge"
Station "London Victoria" (Coach/Bus)3 votes -
Linking the Wikipedia ID to the entities category
The response includes both entities (e.g. Person, Quantity, email and etc.) and Wikipedia ID, but it is only either of them. For example, 'Donald Trump' as input will result in Wikipedia ID 'https://en.wikipedia.org/wiki/Donald_Trump' as output, WITHOUT the entity 'person' which is supported entitiy in the API. Thus, if there any detected Wikipedia ID, could you please add entities type on the response as well? Thank you.
2 votes -
Linkedin company name in English is detected as Portuguese
Request:
POST https://westus.api.cognitive.microsoft.com/text/analytics/v2.0/languages HTTP/1.1
Host: westus.api.cognitive.microsoft.com
Content-Type: application/json
Ocp-Apim-Subscription-Key: f1b41ecadcc64b68b0c5557b27ee770b{
"documents": [{
"id": "1",
"text": "Linkedin"
}]
}RESPONSE:
Transfer-Encoding: chunked
x-ms-transaction-count: 1
x-aml-ta-request-id: 4c95ebec-101e-4e2e-9ec2-5fb61b2153a0
X-Content-Type-Options: nosniff
apim-request-id: f1e4cb45-d93c-4f11-a028-38a0d4a94eda
Strict-Transport-Security: max-age=31536000; includeSubDomains; preload
Date: Tue, 24 Jul 2018 17:59:09 GMT
Content-Type: application/json; charset=utf-8{
"documents": [{"id": "1",
"detectedLanguages": [{
"name": "Portuguese",
"iso6391Name": "pt",
"score": 1.0
}]}],
"errors": []
}1 vote -
Classify the type of entities (people, location, business, dates)
I enjoyed testing the text analytics API, however I noticed that while the service detects entities, it doesn't appear to classify them. For example, I would expect the service to identify a person from a business or street name. Similarly, can it detect an address such as 123 Elm Street, City, State, Zip. I anticipate that this would be a valuable add-on.
6 votes -
Add Wikidata Entity Linking References
The entity linking feature should include or at least allow the option of Wikidata references. Wikidata will allow for much more programmable context vs Wikipedia.
So for example, if the entity "Seattle" was found, we should have a reference to this link from wikidata. "https://www.wikidata.org/wiki/Q5083" which has a ton of contextual data for programmers to access. Such as this is a city, in the state of WA as well as who the mayor is, etc, etc...
1 voteHello,
This feature is in preview. Please see: https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/overview#entity-recognition-preview
Thanks,
Luke -
Entity extraction for fixed format text (ticket number, phone number..etc)
Extract entities from text. example return a Jira task number "TES-10" usually they have a set of rules (ProjectKey-Number) but currently Text analytics is not able to recognize the full entity (returns ProjectKey only).
6 votes
- Don't see your idea?