Topic
  • 3 replies
  • Latest Post - ‏2012-06-12T09:19:46Z by bfoyle
ShivaniP.
ShivaniP.
5 Posts

Pinned topic Sample for IBM Content Analytics

‏2012-06-08T11:10:39Z |
Hi,

I am looking for a some input and corresponding sample outputs for IBM Content Analytics. Note I do not have IBM Content Analytics, so I cannot run anything and explore. Can anyone help on this?

Thanks.
Regards,
Shivani P.
Updated on 2012-06-12T09:19:46Z at 2012-06-12T09:19:46Z by bfoyle
  • bfoyle
    bfoyle
    60 Posts

    Re: Sample for IBM Content Analytics

    ‏2012-06-08T23:15:56Z  
    Can you expand on what you are trying to do?

    Typically, you analyze a bunch of content (documents, websites, email, etc) and then you utilize the Content Analytics Miner interface to identify concepts and insights which will help you understand the content you are analyzing and drive better business outcomes from those insights.

    You can also export those to a database or xml for use in another system or process or to drive business rules.

    ...when you say Are you looking for input and corresponding outputs...are you looking for exports? ...or screenshots of the analytics UI? I'm not sure what you are asking.

    bf
  • ShivaniP.
    ShivaniP.
    5 Posts

    Re: Sample for IBM Content Analytics

    ‏2012-06-11T08:42:24Z  
    • bfoyle
    • ‏2012-06-08T23:15:56Z
    Can you expand on what you are trying to do?

    Typically, you analyze a bunch of content (documents, websites, email, etc) and then you utilize the Content Analytics Miner interface to identify concepts and insights which will help you understand the content you are analyzing and drive better business outcomes from those insights.

    You can also export those to a database or xml for use in another system or process or to drive business rules.

    ...when you say Are you looking for input and corresponding outputs...are you looking for exports? ...or screenshots of the analytics UI? I'm not sure what you are asking.

    bf
    Thanks for your response.
    Consider that we have a huge word documents & they may have tasks and status described in them. Using IBS content analytics, can we extract tasks list & their status? If yes, I want to understand what portion of document is extracted & on what basis?

    Regards, Shivani
  • bfoyle
    bfoyle
    60 Posts

    Re: Sample for IBM Content Analytics

    ‏2012-06-12T09:19:46Z  
    • ShivaniP.
    • ‏2012-06-11T08:42:24Z
    Thanks for your response.
    Consider that we have a huge word documents & they may have tasks and status described in them. Using IBS content analytics, can we extract tasks list & their status? If yes, I want to understand what portion of document is extracted & on what basis?

    Regards, Shivani
    IBM Content Analytics will process the whole document and will extract any concepts or annotations you have specified for a given collection. Typically these annotations are based on prose like these posts and not on structured lists or tables. However, depending on the format of the task list, you might be able to create a UIMA annotator using ICA Studio that would allow it to extract information about the tasks on that list.

    For example(and I'm just making this up since I don't have one of your documents)

    Tasks:
    1. Shivani will post a message in the ICA forum to get more information annotators Status:complete
    2. Bob will respond to Shivani's post Status:complete
    3. Shivani will respond to Bob's post Status:complete
    4. Bob will provide examples Status:in progress

    ICA will, out of the box identify nouns, verbs, adjectives and phrases associated with these tasks..."Bob will respond" for example...or even the word task and numbers. Optionally, you can turn on named entity annotations which will give you things like "Bob" and "Shivani". There are a few other out of the box annotators but those look like the pertinent ones.

    I would suggest to maybe create an custom annotator that looks at an aggregate rule that associates the numbered elements to the task section and then specific actions to each number and then associate the status to each.

    This will all be analyzable in the context of the whole document of sections of the document if you have them defined. So you may be able to analyze things like types of tasks across document clusters or things like that.

    I hope this helps.
    bf