+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 16 to 19 of 19

Thread: Why a simple linear regression (OLS)

  1. #16
    TS Contributor
    Points: 12,227, Level: 72
    Level completed: 45%, Points required for next Level: 223
    rogojel's Avatar
    Location
    I work in Europe, live in Hungary
    Posts
    1,470
    Thanks
    160
    Thanked 332 Times in 312 Posts

    Re: Why a simple linear regression (OLS)




    Ouch,
    I was referring to the first post in the thread and not to you, apologies if this was misunderstabdable. Still, model selection is a HUGE topic, and I very much doubt the the purposeful selection is so much better then the alternatives, that one could recommend it from the start. In the article I qoute they explicitely say that the method is theoretically unproven, supported by a few nunerical simulations only. Also, as much as this is worth , I never heard about it in any MR courses or even books that specifically discuss model selection like here: https://www.amazon.com/Introduction-...8907476&sr=1-2

    So, I guess, hlsmith is right and modern, computer-intensive methods made this obsolete?

    regards

  2. #17
    Omega Contributor
    Points: 38,334, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Why a simple linear regression (OLS)

    I think there was a push saying stepwise model were inferior to statistician built model, now I feel that people are pushing for automated ensemble methods because human bias everything and can be subjective.


    As mentioned, it also comes down to content knowledge related to subject, so I think the two need to be infused.


    Any way, they are all just MODELS.
    Stop cowardice, ban guns!

  3. #18
    Points: 1,741, Level: 24
    Level completed: 41%, Points required for next Level: 59

    Posts
    230
    Thanks
    37
    Thanked 68 Times in 59 Posts

    Re: Why a simple linear regression (OLS)

    Quote Originally Posted by rogojel View Post
    Ouch,
    Sorry if there was any negative tone! I definitely didn't intend that in a combative way (but I suppose that can be hard to convey over text).

    Quote Originally Posted by rogojel View Post
    I was referring to the first post in the thread and not to you, apologies if this was misunderstabdable. Still, model selection is a HUGE topic, and I very much doubt the the purposeful selection is so much better then the alternatives, that one could recommend it from the start. In the article I qoute they explicitely say that the method is theoretically unproven, supported by a few nunerical simulations only. Also, as much as this is worth , I never heard about it in any MR courses or even books that specifically discuss model selection like here: https://www.amazon.com/Introduction-...8907476&sr=1-2

    So, I guess, hlsmith is right and modern, computer-intensive methods made this obsolete?

    regards
    I definitely didn't pick up that it was for the OP, so my apologies again! I think a lot of the approaches with a "system" are often recommended in cases of exploratory analysis to screen potential predictors more or less (absence of prior work or theory). I actually hadn't heard about it in any courses either. My first encounter was in the H&L Applied Logistic Regression book and then again (not surprisingly) in their Applied Survival Analysis book. I had learned some more "traditional" (?) approaches like stepwise methods, building up or down, and some others. Since the H&L book (logistic) is so well respected I figured it was a safer bet (especially since these books were written in a time of good computing resources post-2000).

  4. The Following User Says Thank You to ondansetron For This Useful Post:

    rogojel (03-07-2017)

  5. #19
    Points: 533, Level: 10
    Level completed: 66%, Points required for next Level: 17

    Posts
    50
    Thanks
    20
    Thanked 0 Times in 0 Posts

    Re: Why a simple linear regression (OLS)


    This topic is so active I will check the books and things that are discussed here. Thanks!

    I have one more question though: http://www.talkstats.com/showthread....ple-regression

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats