addressalign-toparrow-leftarrow-rightbackbellblockcalendarcameraccwcheckchevron-downchevron-leftchevron-rightchevron-small-downchevron-small-leftchevron-small-rightchevron-small-upchevron-upcircle-with-checkcircle-with-crosscircle-with-pluscrossdots-three-verticaleditemptyheartexporteye-with-lineeyefacebookfolderfullheartglobegmailgooglegroupshelp-with-circleimageimagesinstagramlinklocation-pinm-swarmSearchmailmessagesminusmoremuplabelShape 3 + Rectangle 1ShapeoutlookpersonJoin Group on CardStartprice-ribbonShapeShapeShapeShapeImported LayersImported LayersImported Layersshieldstartickettrashtriangle-downtriangle-uptwitteruserwarningyahoo

Detecting Very Rare but High Value Events

Prresented by Richard Bijjani, Chief Technology Officer and Co-founder at Quanttus.

When machine learning scientists are confronted with the task of detecting low probability high value events, they find textbook approaches extremely inadequate.

Terrorism and some deadly diseases could be seen as very rare events with disastrous effects. We will discuss some techniques drawn from the security as well as the health care industries that could enable the early detection of ‘yet unseen’ deadly threats.

Join or login to comment.

  • Eve

    Is there a recording of this event available?

    1 · March 6, 2014

  • David L.

    I enjoyed the talk. It was an interesting overview of the problem with examples in different industries.

    NEEDS IMPROVEMENT: the audio volume keep changing due to directional hand-held mic. Perhaps get a lapel microphone for the speaker.

    March 6, 2014

  • TomA

    Speaker started out strong but stayed too general. Didn't share specifics on his approaches to detecting rare events.

    2 · February 26, 2014

    • Scott F.

      Agreed

      February 27, 2014

    • William M. G.

      No magic formula. Know your data, know your data provenence, know the physical model around the data, and don't throw away outliers. Comes with experience, not the latest tool. Each rare event is not like the others so don't expect an easy answer. Very worthwhile talk.

      February 27, 2014

  • Apollo C.

    Amazing event, great presentation, and really interesting chats afterward.

    February 27, 2014

  • Ergest

    For those interested here is a PhD thesis on fraud detection (one type of a rare event) which discusses several methods in detail including smote and under/over sampling. http://www.dataminingmasters.com/uploads/studentProjects/thesis27v12P.pdf

    1 · February 27, 2014

  • Siva G.

    more generic.. i was expecting more details..

    February 27, 2014

  • Theophano M.

    I wasn't able to attend. Would it be possible to post a summary or the presentation itself online?

    February 27, 2014

  • vitaly b.

    I think a speaker should notify in advance about the talk agenda. In this case I am sure most of folks expected something more specific, not just general overview with nice pictures.

    1 · February 27, 2014

  • Nirupam S.

    Very useful

    February 26, 2014

  • Michael F.

    Excellent session

    February 26, 2014

  • Sheamus M.

    I'm waiting on the talk abstract from the speaker and should be receiving it in the next couple of days.

    2 · February 3, 2014

    • scomo

      Hey I'd your still in the areas left my coat behind

      February 26, 2014

  • scomo

    Hey shamus left my coat behind at the meeting

    February 26, 2014

  • Zahir

    I don't think I can make it today

    February 26, 2014

  • Kevin

    may have to drive. is there any parking available nearby? thanks much

    February 26, 2014

  • Pasha

    Can you post more information than these 7 words? Industry, dataset, technology, methods, tools, algorithms?? Is in the context of your consumer health company? Is it a case study or ? We need more information in order to be able to detect a high value event with you this evening.

    1 · February 3, 2014

    • Towne T.

      use algorithms from dataset find business value, and it maybe become a new industry

      February 26, 2014

  • Simon R.

    Sounds interesting

    February 23, 2014

  • Clara

    Is there a more detailed agenda available for this event? When does the talk begin? How long is it?

    February 18, 2014

  • John S.

    Extreme Value Statistics?

    February 13, 2014

  • A former member
    A former member

    What exactly is this talk about? Is it in the context of a specific technology?

    February 3, 2014

    • Perry S.

      Same question here - are these statistical significance or sampling or high volume processing issues? Any detail you could add would be most appreciated.

      February 3, 2014

  • Plamen

    Interesting topic. I am curious to see what approaches people are using.

    February 3, 2014

Our Sponsors

  • Strata + Hadoop World

    20% special discount code UGBOSDM

  • WeWork

    Contact WeWork, mention your BDM Meetup Membership for special discounts

  • Open Data Science

    Signup for ODSC Newsletter. Receive attractive ODSC East Conf. offers!

People in this
Meetup are also in:

Sign up

Meetup members, Log in

By clicking "Sign up" or "Sign up using Facebook", you confirm that you accept our Terms of Service & Privacy Policy