Is there any way to find the row with closest values in matrix with a given array values?

Dear expert,

I have an array (1mat.txt) and I want to check the closest row in the matrix (2mat.txt) with the array. In this 2mat.txt i have 5 rows (which can be seen with graph) but in principle I have bigger matrix like 50 rows or 100 rows. Kindly help me with this.

Thank you

Answer 1

[Torsten]

How do you define "closest" ? The Euclidean distance between array and rows of the matrix ?

Then you can use

mat1 = [1.0000    0.1803    0.0779    0.0475    0.0369    0.0269    0.0208    0.0166    0.0093    0    0.0009    0.0008];
mat2 = [1     0     0     0     0     0     0     0     0     0     0     0
    1.0000    0.1579    0.0244    0.0033         0         0    0.0003         0         0         0         0         0
    1.0000    0.4389    0.1907    0.0842    0.0332    0.0169    0.0061    0.0049    0.0019         0         0         0
    1.0000    0.5587    0.3522    0.2150    0.1275    0.0599    0.0423    0.0275    0.0099    0.0041    0.0038         0
    1.0000    0.5728    0.3655    0.2707    0.1718    0.1064    0.0629    0.0337    0.0184    0.0132         0    0.0010];
[~,i] = min(vecnorm((mat2-mat1).'))

i = 2

[SHUBHAM AGARWAL]

Dear Torsten,

Thank you for the answer. Closest means , If i draw every row of 2mat.txt as graph (y-axis) with x-axis (1:1:12) then which row would be the closest to the graph of 1mat.txt (y-axis) with respect to the x-axis (1:1:12) steps. I am not sure about the method. I hope I am able to clarify.

Thank you

[Torsten]

I am not sure about the method. I hope I am able to clarify.

No. In each point 1:12, the two graphs differ by a certain amount. But what amount do you want to take to say "they are closest" ? The point where the difference is maximum ? The sum of all differences as absolute value ? The sum of the squared differences ? The integral between the two curves ?

[SHUBHAM AGARWAL]

Dear torsten,

The sum of the squared differences.

Thank you

[Torsten]

Then you have the code already - it's my first answer.

[SHUBHAM AGARWAL]

Dear Torsten,

Thanks for the explanation of this method. It works.

Thank you very much.