+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 16 to 30 of 30

Thread: Logistic Regression

  1. #16
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Logistic Regression




    Sorry for any confusion I brought - I realize now it might have sounded like a suggestion. I was questioning noetsi because I don't think it's a good idea.

    Since you have multiple observations at each point you could try calculating the logit of the estimated probabilities and then plotting those against dose. If logistic regression is a good fit then these should be approximately linear.

  2. The Following User Says Thank You to Dason For This Useful Post:

    MrAnon9 (02-08-2012)

  3. #17
    Fortran must die
    Points: 58,790, Level: 100
    Level completed: 0%, Points required for next Level: 0
    noetsi's Avatar
    Posts
    6,532
    Thanks
    692
    Thanked 915 Times in 874 Posts

    Re: Logistic Regression

    Quote Originally Posted by Dason View Post
    ANOVA for binary response?
    A good point. You would have to use one of the non-parametric equivilents. I was thinking of how to address the problem and not the form of the data.
    "Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995

  4. #18
    Points: 2,626, Level: 31
    Level completed: 18%, Points required for next Level: 124

    Location
    Dallas, TX
    Posts
    311
    Thanks
    12
    Thanked 90 Times in 88 Posts

    Re: Logistic Regression

    Quote Originally Posted by Dason View Post
    Since you have multiple observations at each point you could try calculating the logit of the estimated probabilities and then plotting those against dose. If logistic regression is a good fit then these should be approximately linear.
    This sounds like a good solution.

  5. #19
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Logistic Regression

    If there aren't that many observations though it might make sense to smooth the estimated probabilities using some sort bayesian estimator. Something like (x+1)/(n+2) or (x+.5)/(n+1) would probably work well enough.

  6. #20
    Points: 1,278, Level: 19
    Level completed: 78%, Points required for next Level: 22

    Posts
    135
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Re: Logistic Regression

    Hi again, thanks for the help. I have done two plots from the data..

    Are they relevant plots for deciding if the logistic model is suitable and if so, what can I deduce from the two graphs?

    I think for plot 2, it needs to be a straight line relationship? Is this correct and what can I conclude from the first plot?

    Should I plot against dose or log dose, does it matter?

    Thanks in advance guys.
    Attached Thumbnails Attached Thumbnails Click image for larger version

Name:	plot1.jpeg‎
Views:	8
Size:	42.8 KB
ID:	2037   Click image for larger version

Name:	plot2.jpeg‎
Views:	16
Size:	47.7 KB
ID:	2038  

  7. #21
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Logistic Regression

    Why are you plotting the log of the estimated probability in the second plot? We would want a plot of the logit of the estimated probability to analyze if a logistic seems alright.

  8. The Following User Says Thank You to Dason For This Useful Post:

    MrAnon9 (02-09-2012)

  9. #22
    Points: 1,278, Level: 19
    Level completed: 78%, Points required for next Level: 22

    Posts
    135
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Re: Logistic Regression

    I don't get it

  10. #23
    Points: 2,626, Level: 31
    Level completed: 18%, Points required for next Level: 124

    Location
    Dallas, TX
    Posts
    311
    Thanks
    12
    Thanked 90 Times in 88 Posts

    Re: Logistic Regression

    According to the theory, g(proportion)=Xb where g(proportion) is the logit function. Therefore, if your data is a case for logistic function then the logit of the proportion will be linear with predictors i.e. dose. Therefore, Dason asked you to check how does the logit of the proportion(i.e. log(p/1-p) fare when plotted against dose.

  11. The Following User Says Thank You to jrai For This Useful Post:

    MrAnon9 (02-09-2012)

  12. #24
    Points: 1,278, Level: 19
    Level completed: 78%, Points required for next Level: 22

    Posts
    135
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Re: Logistic Regression

    But what is p then? The actual logistic function? That would mean I would have to calculate the coefficients..?

  13. #25
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Logistic Regression

    Logit(x) = Log(x/(1-x)) where Log is the natural log. In your case you would stick the estimated probabilities in for x. So if you had 6 success out of 8 trials you would calculate x= 6/8 = .75. So Logit(.75) = Log(.75/.25)

  14. The Following User Says Thank You to Dason For This Useful Post:

    MrAnon9 (02-09-2012)

  15. #26
    Points: 1,278, Level: 19
    Level completed: 78%, Points required for next Level: 22

    Posts
    135
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Re: Logistic Regression

    Thanks for the quick response. I need to do this by tomorrow :P

    Is that not what I did though? Since log ( ri/ni / 1- ri/ni ) = log (ri / (ni-ri) ) right?

    where ri is number of beetles killed for each dose and ni is the total for each dose, hence ni-ri is the survival

  16. #27
    Points: 1,278, Level: 19
    Level completed: 78%, Points required for next Level: 22

    Posts
    135
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Re: Logistic Regression

    plotted graphs wrong anyway, but still not sure if what i plotted is right (previous reply)

  17. #28
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Logistic Regression

    You would probably want to plot dose on the x axis (instead of log(dose)) but yeah - I didn't put as much thought into it as I probably should have - seems like what you have should work.

  18. #29
    Points: 2,626, Level: 31
    Level completed: 18%, Points required for next Level: 124

    Location
    Dallas, TX
    Posts
    311
    Thanks
    12
    Thanked 90 Times in 88 Posts

    Re: Logistic Regression

    Quote Originally Posted by Dason View Post
    You would probably want to plot dose on the x axis (instead of log(dose)) but yeah - I didn't put as much thought into it as I probably should have - seems like what you have should work.
    I was also earlier thinking that he should plot dose on x axis but then his independent is log dose & not dose. Therefore, it seems what he has done is correct.

  19. #30
    Devorador de queso
    Points: 95,819, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,935
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Logistic Regression


    Quote Originally Posted by jrai View Post
    I was also earlier thinking that he should plot dose on x axis but then his independent is log dose & not dose. Therefore, it seems what he has done is correct.
    Huh. I thought I saw dose as the IV at some point. Looks like I did a whole bunch of screwing up in this thread. Oh well.

+ 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