distance and matrix unique numbers

Could you, guys, help me please. I have a matrix X X =

    2.2000    3.3000
    2.2000    3.3000
    4.0000    5.0000
    3.0000    5.0000
I need to get from this [index,A]
index = [2] or equivalently index = [1] 
and
A =

    2.2000    3.3000
    4.0000    5.0000
    3.0000    5.0000
Then I need to calculate the distance to the nearest neighbor within that matrix A. i.e. it will be 3x1 vector of distances.
Any help please. Thanks

Answer 1

[the cyclist]

I don't really understand the output you want. In the first step, do you just want the unique rows? You can do that with:

 >> [uniqueX,i,j] = unique(X,'rows')

Read the help file for the unique() function for more details.

[jenka]

the cyclist, it works for a simple example above. However, for some reason, it does not work for my bigger problem, i.e. the uniqueX matrix has many rows that are the same (which I want to avoid). Any suggestion?

[jenka]

also, it appears that the first element in UnigueX does not equal the first element in X. Why? Thanks

Answer 2

[jenka]

Hi, well, I tried that but for my data it does not have unique numbers for some reason (maybe it is something to do decimal truncation?) Also, how to then calculate distance vector for matrix uniqueX. Thanks!!!!

Answer 3

[Fangjun Jiang]

Do you mean this:

X=[2.2000    3.3000
    2.2000    3.3000
    4.0000    5.0000
    3.0000    5.0000];
index=2;
A=X;
A(index,:)=[];
D=dist(A')

Function dist() requires Neural Network Toolbox. You may also check the function pdist(), which requires Statistics Toolbox.

Answer 4

[Walter Roberson]

http://matlab.wikia.com/wiki/FAQ#Why_is_0.3_-_0.2_-_0.1_.28or_similar.29_not_equal_to_zero.3F

[jenka]

however, it does give me the same rows. I checked my calculating the nearest neighbor. The results for these "repeated" rows is 0 which is wrong.

[Walter Roberson]

Subtract one of the repeated rows from another: the result will very likely include a non-zero result. Make sure you are using

format long g

as format f might truncate the results.