find the nearest values

Question

Rima Habib 2020-1-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/501769-find-the-nearest-values

评论： Rima Habib 2020-1-24

采纳的回答： Stephen23

在 MATLAB Online 中打开

Hallo,

I have a matrix B (5000*2) and A(1*2)

now I need to find the nearest values in B to A

meaning the closest nearest value in the first column of B to the A(1,1), and the closest nearest value in the second column of B to A(1,2)

the 2 values must have the same index, meaning the closet values must be taken as a couple not every one from different indexes.

I tried some thing like :

dist2 = sum((A - B) .^ 2, 2);
closest = B(dist2 == min(dist2),:);

but it only gives me the closest value of the first column and not the both.

and when I search for the values, I could find a better fit but its really hard as I need to to this for many points not just this one.

Does any one have any suggestions?

Thanks ahead!

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Stephen23 2020-1-24

2
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/501769-find-the-nearest-values#answer_411784

编辑：Stephen23 2020-1-24

在 MATLAB Online 中打开

You were almost there, you just need to use the second output from min to make it easier:

>> A = rand(1,2)
A =
   0.39011   0.53114
>> B = rand(500,2);
>> [D,X] = min(sum((A-B).^2,2));
>> B(X,:)
ans =
   0.41172   0.53908

If A has multiple rows then you can use permute like this:

>> A = rand(3,2)
A =
   0.26960   0.78802
   0.47827   0.75230
   0.34524   0.46480
>> B = rand(500,2);
>> [D,X] = min(sum((A-permute(B,[3,2,1])).^2,2),[],3);
>> B(X,:)
ans =
   0.26950   0.77584
   0.49448   0.76722
   0.34736   0.48158

Note that both of these examples require the scalar dimension expansion supported by MATLAB R2016b or later, for earlier MATLAB versions you will need to replace the subtraction with bsxfun.