KDnuggets Home » Polls » Privacy-preserving data mining via randomization (Aug 2002)

Privacy-preserving data mining via randomization


 
  
Poll
Using privacy-preserving data mining via randomization will make consumers to give more truthful responses [50 votes total]

More frequently (17) 34%
About the same (27) 54%
Less frequently (1) 2%
Don't know (5) 10%

Comments


  • Chris Clifton, Privacy != randomization
    Data perturbation, or randomizing the data, isn't the only way to achieve privacy while still allowing data mining. Even if consumers don't accept the randomization approach, it doesn't mean we shouldn't keep looking for approaches they will accept.

  • Peter, Privacy is not important - what's in it for me is!

    Loss of privacy is just one of the costs associated with answering a question, in addition to more important costs like time invested.

    Frankly, most customers don't care that much about the (privacy) costs, but more about the value: what's in it for me if I answer this question.

    This is the most important issue to be adressed.

  • Editor, Privacy-Preserving Data Mining
    Recent work by IBM researchers -- see
    http://www.kdnuggets.com/news/2002/n14/21i.html
    showed that data mining can preserve privacy by randomizing individual answers yet preserve overall patterns. Some think that programs like can lead to greater truthfulness in the answers people provide on the Web.

    What do you think?

KDnuggets Home » Polls » Privacy-preserving data mining via randomization (Aug 2002)