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-lineeyefacebookfolderfullheartglobegmailgooglegroupsimageimagesinstagramlinklocation-pinm-swarmSearchmailmessagesminusmoremuplabelShape 3 + Rectangle 1outlookpersonJoin Group on CardStartprice-ribbonImported LayersImported LayersImported Layersshieldstartickettrashtriangle-downtriangle-uptwitteruseryahoo

How to be smooth

Ross Gayler describes himself as a recovered psychologist who studied rats and stats (minus the rats) a very long time ago.Since then he has mostly worked in credit scoring (predictive modelling of customer behaviour in retail finance) and has forgotten most of the statistics he ever knew.However, he is very interested in avoiding the yawning chasms that data providers like to add to their data.Consequently, he tries to examine his data very closely using techniques that tend not to hide the data blemishes by assuming restrictive functional forms or arbitrarily aggregating values to a coarse partition.

One of his favourite techniques is nonparametric scatterplot smoothing (a method for modelling a dependent variable as a smooth function of small number of independent variables).In this talk, Ross will cover the background of scatterplot smoothing, and the locfit package (which is probably the most useful implementation of this technique in R).He will also touch briefly on related packages such as gam and mgcv, andstats package functions such as lowessloess, and supsmu.

Join or login to comment.

  • Peter W.

    Enjoyed the presentation. Nice introduction to LOWESS and credit scoring.

    April 26, 2011

  • Xiaonan J.

    Thank you all for the kindly response. I will make sure arrive earlier next time.

    April 20, 2011

  • Michael C.

    Interesting subject matter, some of it over my math level, but relevant.

    April 20, 2011

  • Tony L.

    Very interesting and useful. I'd used many of these functions before but didn't know about locfit

    April 20, 2011

  • Carol L.

    It was excellent and extremely relevant. I think these techniques should be more widely used for scorecard scaling, for variable interaction and local linear modelling for more granular scores.

    April 20, 2011

  • Yuval M.

    Hi Xiaonan, that's unfortunate. Sorry you were not able to make it inside. Yes, we do know that public access closes at 6:30, and in fact if you read the venue information I've put a note on that there. In the past I've also mentioned it in the meetup announcements, but I'm afraid I'm not going to remember to always do that.

    April 20, 2011

  • Yuval M.

    Who would have thought that scatterplots could be so stimulating?? Thanks Ross for introducing me to the scatterplot-smoothing technique - I wasn't aware of it! Great to see an interesting technique being applied to practical business problems.

    April 20, 2011

  • Ross G.

    That must have been really frustrating for you. I presume that Yuval and Jeromy weren't even aware that the lifts locked after 6.30. Now that we know I'm sure they'll put some instructions in future Meetup announcements.

    April 19, 2011

  • Jeromy A.

    I enjoyed Ross' presentation.
    It was good to hear about the real world issues associated with applications in credit scoring.
    It was also helpful to see how R could be applied in this domain.

    April 19, 2011

  • Gad A.

    Great talk - technical enough but not distracted by technicalities. Good discussion too.

    April 19, 2011

  • Xiaonan J.

    Hi Ross, I typed in 11 on the keypad and it asks for access code/pass, which I don't have. There was a cleaner kindly told me to call the security from the helpdesk. The security asked me to ring someone on level 11 to come down and let me in. Obviously I didn't know anyone's phone number at that time. So have to give up.

    I tried all of the eight lifts at around 6.30pm and all the keypads asked for access code/pass.

    April 19, 2011

  • Ross G.

    I find the lifts at Deloitte confusing and didn't know how to use them the first time. You have to enter your destination floor on the keypad in the lobby and it tells you which lift to wait for. Once you are in the lift you can't tell it what floor you want.

    Are you saying that the keypad would not accept a request for the 11th floor or that you didn't know what you were supposed to do?

    April 19, 2011

  • Xiaonan J.

    I arrived at 6:30. Yes it was a bit late but unfortunately there was no one to let me get in to the lift. So I left with disappointment. Could we sort out a nicer way to allow late-arrived people in? Maybe it is not a problem at all because you shouldn't be late in the first place. I really don't know.

    April 19, 2011

Our Sponsors

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