Microsoft

Assaf Israel [MSFT]

My feedback

  1. 1 vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Resolved  ·  1 comment  ·  Text Analytics » Sentiment Analysis  ·  Flag idea as inappropriate…  ·  Admin →
    An error occurred while saving the comment
    Assaf Israel [MSFT] commented  · 

    Thank you for the feedback!
    We recently released a new sentiment model in preview which should improve on such cases. It also provides a breakdown of the sentiment by sentence.

    https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/how-tos/text-analytics-how-to-sentiment-analysis#sentiment-analysis-v3-public-preview

  2. 1 vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    1 comment  ·  Text Analytics » Entity Recognition  ·  Flag idea as inappropriate…  ·  Admin →
    An error occurred while saving the comment
    Assaf Israel [MSFT] commented  · 

    The bingId field can be used in conjunction with the Bing Entity Search API to extract additional information about the returned entities.

    For example, given the following sentence: "I'm going to Yosemite this summer." the response from the entities endpoint will contain the following entry:

    {
    "name": "Yosemite National Park",
    "matches": [
    {
    "text": "Yosemite",
    "offset": 13,
    "length": 8
    }
    ],
    "wikipediaLanguage": "en",
    "wikipediaId": "Yosemite National Park",
    "wikipediaUrl": "https://en.wikipedia.org/wiki/Yosemite_National_Park",
    "bingId": "3b9c9bbf-291d-57e3-a3d5-d563b0f0aadb",
    "type": "Location",
    "subType": null
    }

    You can now use the entity search API to receive additional information about the entity programmatically by submitting the entity name and use the bingId to resolve disambiguations:

    {
    ...
    "description": "Yosemite National Park is an American national park located in the western Sierra Nevada of Central California, bounded on the southeast by Sierra National Forest and on the northwest by Stanislaus National Forest. The park is managed by the National Park Service and covers an area of 748,436 acres and sits in four counties: centered in Tuolumne and Mariposa, extending north and east to Mono and south to Madera County. Designated a World Heritage site in 1984, Yosemite is internationally recognized for its granite cliffs, waterfalls, clear streams, giant sequoia groves, lakes, mountains, meadows, glaciers, and biological diversity. Almost 95% of the park is designated wilderness.",
    "bingId": "3b9c9bbf-291d-57e3-a3d5-d563b0f0aadb",
    "webSearchUrl": "https://www.bing.com/entityexplore?q=Yosemite+National+Park&filters=sid:%223b9c9bbf-291d-57e3-a3d5-d563b0f0aadb%22&elv=AXXfrEiqqD9r3GuelwApulp4JqFUrZucb7ThuhON8*FVPSZDhpCdOcES7WGcz1FBjJNHqEnaopAOhZTEye4PNJucBh!WSqCXrO!fCuflAMDi",
    "name": "Yosemite National Park",
    "url": "https://www.nps.gov/yose/index.htm",
    ...
    }

  3. 2 votes
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    1 comment  ·  Text Analytics » Sentiment Analysis  ·  Flag idea as inappropriate…  ·  Admin →
    An error occurred while saving the comment
    Assaf Israel [MSFT] commented  · 

    We have recently released a new version of sentiment analysis.
    You can read more about the new v3.0-preview API in the following link - https://docs.microsoft.com/en-us/azure/cognitive-services/text-analytics/how-tos/text-analytics-how-to-sentiment-analysis#sentiment-analysis-v3-public-preview

    The new API will consider most of these kind of vertical specific sentences as factual and thus Neutral.
    If you are interested in domain specific flavors of sentiment analysis feel free to open a new suggestion.

  4. 1 vote
    Sign in
    (thinking…)
    Sign in with: Facebook Google
    Signed in as (Sign out)

    We’ll send you updates on this idea

    Under Review  ·  1 comment  ·  Text Analytics » Sentiment Analysis  ·  Flag idea as inappropriate…  ·  Admin →
    An error occurred while saving the comment
    Assaf Israel [MSFT] commented  · 

    Note that this is a warning, not an error. Your document was still processed, but it was scored based on the first 100 tokens of the problematic sentence.

    This error happens when the input contains a sentence with over 100 tokens (words) which usually suggest the data does not contain natural language text (e.g. encoding issues, HTML or binary content, etc).

Feedback and Knowledge Base