Page 134 - Artificial Intellegence_v2.0_Class_12
P. 134

p
                                                              y  – y )
                                                                  i
                                                              i
                                                    MSE    i=i
                                                               n
        MSE will never be negative because the errors are always squared.
        A dvantag e
            is useful for ensuring that our trained model does not have any outlier predictions  ith significant errors because
            places a higher  eight on these errors due to the squaring element of the function.

        Disadvantag e
        If our model makes a single particularly incorrect forecast, the squaring part of the function multiplies the error.  o ever,
        in many real life cases,  e don't  orry about these outliers and instead seek a more fully rounded model that performs
         ell enough on ma ority of cases.
                                                                                             Experiential Learning

                    Video Session


                can the    code or visit the follo ing link to  atch the video   ean  quared  rror
               https      .youtube.com  atch v  h      ma
               After  atching the video, ans er the follo ing question
                hat do you mean by









         or calculating     in  ython, the mean squared error function gives the  ean squared error regression loss.

        f r om sk lear n.metr ics imp or t mean_ sq u ar ed _ er r or
        y _ tr u e =  [ 3 ,   0 .5 ,   ,   .2 ]     list of actual values
                               2
                        -
                                   7
        y _ p r ed  =  [ 2 .5 ,   .0 ,   .3 ,   ]    list of predicted values
                          0
                                      8
                                2
        p r int( " M SE  v alu e= " , mean_ sq u ar ed _ er r or ( y _ tr u e,  y _ p r ed ) )   returns     value
          sing nested lists
        y _ tr u e =  [ [ 0 .5 ,   ] , [ - 1 ,   ] , [ 7 ,   5 ] ]
                                             -
                                     0
                           1
                                             -
                                   1
        y _ p r ed  =  [ [ 0 ,   ] , [ - 1 ,   .5 ] , [ 8 ,   5 .5 ] ]
                         2
        p r int( " M SE  v alu e= " , mean_ sq u ar ed _ er r or ( y _ tr u e,  y _ p r ed ) )
        Output:
            value   .
            value   .
        R MS E ( R oot Mean S q uar e Er r or )
         he  oot  ean  quare  rror      ) is a metric for determining ho   ell a regression line fits the data points.
        can also be understood as the standard deviation in the residuals.  ecall that the residual is the difference bet een the
        predicted value and observed value in the  egression  ine.






                      Touchpad Artificial Intelligence (Ver. 2.0)-XII
   129   130   131   132   133   134   135   136   137   138   139