1. 07-01-11 11:41 AM #1
    luegofuego
    luegofuego's Avatar Become A Pro!
    Join Date: 06-16-10
    Posts: 96
    Betpoints: 12

    Just getting started with regressions

    I've acquired a database which I'm dicking around with, trying to learn as much as I can about modelling some stuff.

    Finding linear correlations isn't that hard, just click some buttons in SPSS and there you go - but I'm still not sure how to model binary dependant variables - such as finding out the likelihood of a team WINNING instead of calculating what the average point difference will be.

    I tried formating it as a dummy variable (0 for no, 1 for yes) but that spat out pretty damn weird numbers. I'm guessing there's a better tool than standard linear regressions here?

    Appreciate the help
    Reply With Quote Reply With Quote
  2. 07-01-11 12:25 PM #2
    bztips
    bztips's Avatar Become A Pro!
    Join Date: 06-03-10
    Posts: 283
    Betpoints: 6

    Logit (or probit) regression -- I'm sure SPSS can do those.
    Reply With Quote Reply With Quote
  3. 07-03-11 03:16 AM #3
    brettd
    brettd's Avatar Become A Pro!
    Join Date: 01-25-10
    Posts: 229
    Betpoints: 3869

    Yeah logit regression. Jump on you tube and take a look. Just finished logistic regression in my course. Fairly simple stuff.
    Reply With Quote Reply With Quote
  4. 07-03-11 01:29 PM #4
    Peregrine Stoop
    Peregrine Stoop's Avatar Become A Pro!
    Join Date: 10-23-09
    Posts: 869
    Betpoints: 779

    Khan Academy videos are usually more helpful than youtube videos for statistics, probability, and mathematics
    Reply With Quote Reply With Quote
  5. 07-03-11 10:22 PM #5
    Insoluble
    Insoluble's Avatar Become A Pro!
    Join Date: 05-29-10
    Posts: 71
    Betpoints: 3463

    Agreed, Khan academy can be a great resource for learning a ton about anything
    Reply With Quote Reply With Quote
Top
« Previous Thread | Handicapper Think Tank | Next Thread »