Hello everyone. I got some data (x=εs y=Uslip z=dcl or φ), and plotted them in a 3D scatter plot, see below
and x-z, y-z, x-y plots are also given as
according to dcl-uslip , it seems that the data can be classified into 2 groups, so I wanna fit 2 surfaces in one plot, but I dont know how to do this.
can anybody give me some help?
thx
Solved! Go to Solution.
Your data is defined as x, y, and z vectors, so you only need a vector of residuals, not a matrix.
Something like the attached? There are ways to automatically separate data into groups (look up "cluster analysis") but this case was so easy I just defined a line by eye.
Thank you very much! Richard
your suggestion works finely to my other data sets
and after splitting the data I did some regressions to the raw data sets using the procedure recommonded by mathcad (see following), and extracted the coefficients
and when I tried fitting the data by a user-defined objective function with minerr function , dimension matchness problems and strange results were always encountered
does anybody know how do I avoid the problem? or is it correct that I use minerr function in such a case?
Your data is defined as x, y, and z vectors, so you only need a vector of residuals, not a matrix.
Oh!!
Many Thanks
YOU ARE RIGHT! I defined data as vectors, but used a matrix to calculate the residuals. It is too late last midnight, I m too sleepy to find this.
Thank you again for your answer!