Sign in
 
Home20122008 R2200820052000LibraryForums
SQL Server >
> 
> 
RMSE Calculation for Discrete Targets

Answered RMSE Calculation for Discrete Targets

  • Friday, January 18, 2013 6:50 AM
     
     
    Sign In to Vote
    0

     

    I was just wondering if anyone knows how the Root Mean Square Error (RMSE) is calculated for Discrete Targets? 

    Is a correct prediction calculated as 1 and so an incorrect prediction would be 0? Or is there a little more to it than this?

    Thanks in advance.

     

All Replies

  • Monday, January 28, 2013 10:02 AM
    Moderator
     
     Answered
    Sign In to Vote
    0

    Hi Liz Wu,

    Which measure is the Root Mean Square Error in?

    Likelihood: Square root of the mean error for all partition cases, divided by the number of cases in the partition, excluding rows that have missing values for the target attribute. MSE is a popular estimator for predictive models. The score averages the residuals for each case to yield a single indicator of model error.

    Estimation: Average error when the predicted value is compared to the actual value.RMSE is a popular estimator for predictive models. The score averages the residuals for each case to yield a single indicator of model error.

    For more information about it, please see: http://msdn.microsoft.com/en-us/library/bb895177.aspx

    Thanks,
    Eileen

    Eileen Zhao
    TechNet Community Support