That's well posible, it's called mean absolute deviation instead of variance. There are some reasons why the errors or deviations in Statistics traditionally got squared. When you need to estimate a quantity, which is extremely common in Statistics, a general procedure is to consider the error in the estimation and choose the candidate value which minimizes that error. You may know that derivatives are a powerful tool for finding a minimum, since the derivative at a minimum must be 0. The square function is derivable, so this simple technique is available for finding the explicit formula of the value which will minimize the error, while the absolute value is not derivable. This is an often mentioned reason for choosing to square things. However, by squaring errors/deviations one magnifies greatly the influence of data points with big errors/deviations, which may be pretty negative sometimes. Since the 1960s, a lot of work has been done in Robust Statistics, which attempt to limit the influence of a few data behaving substantially different from all the others. In this context the mean absolute deviation is better than the variance, and the median is better than the mean. Hope this was not overlong :)
hey Eddie, thank you so much now I can understand what I'm actually doing with the formulas, you're a great teacher, love from Malaysia
I can't believe i actually understood that
This was dope
Very useful
Good
what grade do you teach??
Hi, I want to translate Arabic, please... 🇮🇶Add the Arabic language violin, I am following you from Iraq
Why not take the average distance from the mean using absolute value instead of getting the squared distance.
That's well posible, it's called mean absolute deviation instead of variance. There are some reasons why the errors or deviations in Statistics traditionally got squared. When you need to estimate a quantity, which is extremely common in Statistics, a general procedure is to consider the error in the estimation and choose the candidate value which minimizes that error. You may know that derivatives are a powerful tool for finding a minimum, since the derivative at a minimum must be 0. The square function is derivable, so this simple technique is available for finding the explicit formula of the value which will minimize the error, while the absolute value is not derivable. This is an often mentioned reason for choosing to square things.
However, by squaring errors/deviations one magnifies greatly the influence of data points with big errors/deviations, which may be pretty negative sometimes. Since the 1960s, a lot of work has been done in Robust Statistics, which attempt to limit the influence of a few data behaving substantially different from all the others. In this context the mean absolute deviation is better than the variance, and the median is better than the mean.
Hope this was not overlong :)
@@pedroteran5885 I love you Pedro you taught me stuff there.... Thanks lad!
9:33 clapping, huh? I'd clap if I'd Eddie for a teacher
Yummie yummie poo poo come fresh from eddy woo woo