Ask HN: Are there examples of sentiment analysis in a professional context?
I've never seen sentiment analysis that was convincing in a professional, business context.
Understanding the context of sentences is challenging (e.g. is it a question or an opinion? Which part of the sentence is relevant for the sentiment?).
1 comments
Yes, it is used to power many tech companies chat and text comprehension tools and services (e.g. Alexa, Siri, Hey Google, AWS Comprehend, Google Cloud Natural Language API, Azure Text Analytics, IBM Watson Natural Language Understanding, etc.)
With this customers feed the service the data or have humans or machines interact with the service and normally either conversations, metrics/analytics are derived from those services.
Example, a video might be uploaded to the service of someone that does not speak the language in the video and it has no captions, that service can then process the audio in the video and translate the text and depending on the tech do audio overlay and captions of that video to the language of the person that does not speak the original language in the video.
This could also expand into breakdowns of what was said in the video to translate into thousands of languages, pull in research data from the video, process and build a graph or other infomatics all based on what was said and seen in the video to auto determine and pull in related information based on how things were said and build on what was learned over time.
Another one could be analysis of a person talking to their infotainment system and sharing the results not only based on what they originally said, but also based on the tone and other things going on in the background to differentiate the tone and any additional context of the response. This could lead to a full on conversation that seems natural but with the information gathered allow that system to do a large amount of planning, ordering and other background tasks to accommodate the requests of the user for large planning items (wedding, property management, taxes, accounting, security, plane trips, etc.).
With this customers feed the service the data or have humans or machines interact with the service and normally either conversations, metrics/analytics are derived from those services.
Example, a video might be uploaded to the service of someone that does not speak the language in the video and it has no captions, that service can then process the audio in the video and translate the text and depending on the tech do audio overlay and captions of that video to the language of the person that does not speak the original language in the video.
This could also expand into breakdowns of what was said in the video to translate into thousands of languages, pull in research data from the video, process and build a graph or other infomatics all based on what was said and seen in the video to auto determine and pull in related information based on how things were said and build on what was learned over time.
Another one could be analysis of a person talking to their infotainment system and sharing the results not only based on what they originally said, but also based on the tone and other things going on in the background to differentiate the tone and any additional context of the response. This could lead to a full on conversation that seems natural but with the information gathered allow that system to do a large amount of planning, ordering and other background tasks to accommodate the requests of the user for large planning items (wedding, property management, taxes, accounting, security, plane trips, etc.).