+ Reply to Thread
Page 3 of 3 FirstFirst 1 2 3
Results 31 to 34 of 34

Thread: Comparing coefficients from two separate regressions

  1. #31
    Human
    Points: 12,676, Level: 73
    Level completed: 57%, Points required for next Level: 174
    Awards:
    Master Tagger
    GretaGarbo's Avatar
    Posts
    1,362
    Thanks
    455
    Thanked 462 Times in 402 Posts

    Re: Comparing coefficients from two separate regressions




    Of course multicolinearity influences the evaluations of individual parameter estimates. But how does it influence a Chow test?

    (By the way: MC is often an abbreviation for Markov Chain or Monte Carlo.)

  2. #32
    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: Comparing coefficients from two separate regressions

    I doesn't as far as I know. I never commented on MC in regards to a chow test.

    Pretty much everyone in analsys uses their own nomeclature. For instance SEM stand to some for structural equation models and to others standard error of the measurment. I doubt the OP thought I was talking about either monte carlo simulations or markov chains - which never came into this discussion.
    "Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995

  3. #33
    Human
    Points: 12,676, Level: 73
    Level completed: 57%, Points required for next Level: 174
    Awards:
    Master Tagger
    GretaGarbo's Avatar
    Posts
    1,362
    Thanks
    455
    Thanked 462 Times in 402 Posts

    Re: Comparing coefficients from two separate regressions

    Did Shakespeare write a play about 'much ado about nothing'?

    I wish the original poster good luck if it happens that she returns to this thread.

  4. #34
    Points: 254, Level: 5
    Level completed: 8%, Points required for next Level: 46

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Comparing coefficients from two separate regressions


    Reviving this thread...

    I'm evaluating two (2) refrigerants (gases) that were used in the same refrigeration system.

    I have saturated suction temperature (S), condensing temperature (D), and amperage (Y) data for the evaluation.
    There are two (2) sets of data; 1st refrigerant (R1) & 2nd refrigerant (R2).

    I'm using a non-linear, multivariate (S & D), model for the regression analyses;
    polynomial of the 3rd degree.

    I would like determine how much less/more amperage on average, as a percentage, is being drawn by the second refrigerant.
    Or, some similar metric as a performance comparison.


    First thought was :
    1. Determine the model to use...
    Y = b0 + b1*S + b2*D + b3*S*D + b4*S^2 + b5*D^2 + b6*S^2*D + b7*D^2*S + b8*D^3 + b9*S^3
    2. Derive coefficients (b) from the baseline data (R1).
    3. Using those coefficients, for each S & D in the R2 data set, calculate each expected amp draw (Y-hat) and then average.
    4. Compare that Y-hat average to the actual average amp draw (Y2) of the R2 data.
    5. percent (%) change = (Y2 - Y-hat) / Y-hat

    However, since the 2nd refrigerant has slightly different thermal properties & small changes were made to the refrigeration system (TXV & superheat adjustments) I don't believe this 'baseline comparison method' is accurate.


    Next thought was to do two (2) separate regression analyses:
    Y1 = a0 + a1*S1 + a2*D1 + a3*S1*D1 + a4*S1^2 + a5*D1^2 + a6*S1^2*D1 + a7*D1^2*S1 + a8*D1^3 + a9*S1^3
    &
    Y2 = b0 + b1*S2 + b2*D2 + b3*S2*D2 + b4*S2^2 + b5*D2^2 + b6*S2^2*D2 + b7*D2^2*S2 + b8*D2^3 + b9*S2^3

    and then, for saturated suction temp (S), compare coefficients (a1 vs b1)...
    % change = (b1 - a1) / a1

    However, again, these coefficients should be weighted differently...
    Therefore, the results would be skewed.


    I believe I could use a Z-test to determine how differently weighted the coefficients are, but I'm not sure I fully understand the process, or meaning of the output.
    Z = (a1 - b1) / sqrt ( SE[a1]^2 + SE[b1]^2 )
    But, that still wouldn't give me a performance metric, which is the overall objective...
    And, I believe the Z-test can't be used on non-linear regression models either, correct?


    There's also the option of adding a dummy variable (interacting variable), but I'm not so familiar with how to interpret a performance metric from the regression results...

    Help?!

    Thanks in advance,
    Last edited by gth826a; 02-19-2014 at 03:15 PM. Reason: variable nomenclature update
    ~ "...and speak to me as if I were a small child."

+ Reply to Thread
Page 3 of 3 FirstFirst 1 2 3

           




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