+ Reply to Thread
Results 1 to 2 of 2

Thread: MLR: R2 and RMSE in assessing quality

  1. #1
    Points: 4,633, Level: 43
    Level completed: 42%, Points required for next Level: 117

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    MLR: R2 and RMSE in assessing quality




    How would one judge the quality of a multiple linear regression model if the model had a high value for cross-validated R2 but also a high value for RMSE? How much, or how little, faith should one have that the model could be used to make reliable predictions on future observations? Thanks, in advance, for your response.

  2. #2
    TS Contributor
    Points: 17,636, Level: 84
    Level completed: 58%, Points required for next Level: 214
    JohnM's Avatar
    Posts
    1,948
    Thanks
    0
    Thanked 6 Times in 5 Posts

    The RMSE probably tells you more about the errors on prediction.

    Even though a model may have a high R^2, it's really only one side of the story, and it boils down to how well it predicts actual values.

    The standard error of prediction is probably the best measure of the model.

+ Reply to Thread

           




Similar Threads

  1. Rmse
    By rafters in forum Regression Analysis
    Replies: 0
    Last Post: 11-24-2010, 06:07 PM
  2. ANCOVA for assessing mediators of therapeutic outcome?
    By chrismanchester in forum Psychology Statistics
    Replies: 0
    Last Post: 08-03-2010, 07:08 AM
  3. Linear regression for assessing predictors of variance
    By dgriesdale in forum Psychology Statistics
    Replies: 0
    Last Post: 11-23-2008, 11:55 PM
  4. RMSE calculation related
    By re444 in forum Psychology Statistics
    Replies: 3
    Last Post: 10-26-2008, 06:31 PM
  5. Evaluating RMSE for MLR models
    By Grateful in forum Statistics
    Replies: 1
    Last Post: 04-27-2006, 11:49 AM

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