How to calculate absoulte error?

Question

Sam 2015-1-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/170044-how-to-calculate-absoulte-error

回答： Image Analyst 2015-1-15

absolute_error.png

I've got a datavector (775X1) (look attachment). How do I calculate the absolute error of it? Thanks

2 个评论
显示无隐藏无

John Petersen 2015-1-15

Typically we compute an error by using a truth with estimated values or desired with a fedback value. What does the data vector you have provided represent? If it's the error, then absolute error could be just the absolute value of the vector. Or if you want to represent the error with a scalar then you could take root mean square. or average the absolute value. It depends on what you want.

Sam 2015-1-15

this vector represents the force measured on a forceplate, in Newton.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Image Analyst 2015-1-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/170044-how-to-calculate-absoulte-error#answer_164982

Like John said, you need to know the "true" values to know the error. Or at least some educated guess at what they might be. Why do you think there is any error at all ? Maybe there's not. If you think there is , then you must have some idea of what the values should be. Given some true, or at least reference, values you can use metrics like PSNR or MAD or RMS to quantify the error or residuals.