Matrix compare and unkonw relationships to find
1 回表示 (過去 30 日間)
古いコメントを表示
I have 2 Matrix, for example
Matrix A ={a1 b1 c1; a2 b2 c2; a3 b3 c3}
Matrix B={x1 y1 z1 m1 n1; x2 y2 z2 m2 n2; x3 y3 z3 m3 n3 }
a1 b1 c1 is the result from FEM-Simulation-1, and x1 y1 z1 m1 n1 is the Geometrie-parameter from Simulation-Objekt-1
a2 b2 c2; x2 y2 z2 m2 n2 and the other number in Matrix means the same thing from Simulation-2 and -3.
there must some relationships zwischen Matrix A and B, linear or unlinear, but i don't yet.
How kann i use Matlab this unkonw relationships to find?
Or is it possible to use Matlab this unkonw relationships to find?
which function should i use?
Thankyou very very much for your time!
0 件のコメント
採用された回答
Sourabh Kondapaka
2020 年 8 月 6 日
Hi,
You can use “corr()” function to find the relationship between columns of 2 matrices. Here, as you want to find the correlation between rows, you can transpose both the matrices.
Consider two random matrices of sizes 3x3 and 3x5 respectively
matrix_A = rand(3,3);
matrix_B = rand(3,5);
correlationMatrix = corr(matrix_A', matrix_B' );
3 件のコメント
Sourabh Kondapaka
2020 年 8 月 10 日
Hi,
corr(matrix_A, matrix_B) function returns a matrix of the pairwise linear or rank correlation coefficient between each pair of columns in the input matrices matrix_A and matrix_B.
その他の回答 (1 件)
Bruno Luong
2020 年 8 月 7 日
編集済み: Bruno Luong
2020 年 8 月 7 日
Use regression methods. If you have a linear/affine model, then use linear algebra. If you have "kind" non-linear, use polynomial, spline, fraction, nurbs regressions, if you have no clue on non-linearity but know the relationship is continuous/C1 relationship use learning technique, neuronal netwroks, deeplearning etc... If you have discontinuous non-linearity (hash code, encryption), you might be in big trouble or wait until a real quatum computer available...
参考
カテゴリ
Help Center および File Exchange で Interpolation についてさらに検索
製品
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!