+ Reply to Thread
Results 1 to 4 of 4

Thread: Regression Model is not full rank

  1. #1
    Points: 4,639, Level: 43
    Level completed: 45%, Points required for next Level: 111

    Posts
    46
    Thanks
    12
    Thanked 4 Times in 4 Posts

    Regression Model is not full rank




    Hi all,

    I am trying to detect relationship of crop yield at the end of the growing seasson (1 dependent variable) using amount of precipitation in May, June, July, May+June, June+july, Temperature in May etc. The problem is that I can't get p-values for many of these variables because of the following messages:

    Note: Model is not full rank. Least-squares solutions for the parameters are not unique. Some statistics will be misleading. A reported DF of 0 or B means that the estimate is biased.

    Note: The following parameters have been set to 0, since the variables are a linear combination of other variables as shown.

    precJune = 0.37406 * precH - 4.95972 * avairtH - 0.01172 * precMay


    So, does anyone knows how to solve the problem? If there is another way to analyze the data, apart from regresion please let me know!

    Thanks

  2. #2
    Devorador de queso
    Points: 95,540, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    Posting AwardCommunity AwardDiscussion EnderFrequent Poster
    Dason's Avatar
    Location
    Tampa, FL
    Posts
    12,930
    Thanks
    307
    Thanked 2,629 Times in 2,245 Posts

    Re: Regression Model is not full rank

    If you have the amount of precipitation in May and June in your model why are you trying to add the amount in May+June into the model when that is fully covered by May and June separately? If May+June is literally just the May and June amounts added together you get nothing extra from adding that predictor and that is what your software is complaining about.
    I don't have emotions and sometimes that makes me very sad.

  3. #3
    Points: 4,639, Level: 43
    Level completed: 45%, Points required for next Level: 111

    Posts
    46
    Thanks
    12
    Thanked 4 Times in 4 Posts

    Re: Regression Model is not full rank

    Dason,
    Thanks for the reply.
    I should mention that I tried using the individual months weather data and SAS still finds relationships between my independent variables...

    I extracted some information running proc corr but I would like to run regression and quantify any possible relationship..
    Do you think I should run several simple linear regressions using only 1 independent variable each time?

    Thanks again!

  4. #4
    Points: 2, Level: 1
    Level completed: 3%, Points required for next Level: 48

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Regression Model is not full rank


    hi, how did you resolve this issue. Even i am facing this issue.

+ Reply to Thread

           




Tags for this Thread

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