How to calculate the distance between one row vector and a matrix??...

I need to calculate the distance between a one row vector a and a matrix i.e.

 test_point = [ 1.2 0.6 3.1 1.3]

 train_data = [7.7 2.8 6.7 2.3;
               5.2 2.5 3.0 1.1;
               2.0 6.5 8.6 9.8; 
               1.3 2.4 5.2 3.6;
               4 3 5 4 2 6 7 6]

I have to find the distance of test_point with every row of train_data.... how can I do that??.. I have tried the MATLAB functions pdist2() and dist() and many other etc but they required two matrices of the same dimension...

plzz tell how to write the function to calculate the distance??...

Answer 1

[Andrei Bobrov]

[edit]

a = bsxfun(@minus,train_data,test_point);
out = cellfun(@norm,num2cell(a,2));

[sani ars]

I told you before it gives me the same error as

 Error using pdist2 (line 133)
 X and Y must have the same number of columns.

I have tried your given line of code..... the same error is there...

[Andrei Bobrov]

I am corrected.

[Teja Muppirala]

You didn't need to transpose it, this works fine:

out = pdist2(train_data,test_point)

[sani ars]

What about "Euclidean distance" if I want to use it??... pdist2() is giving me the choices to use any distance calculation mathod i.e "Euclidean distance" ... but with your provided function bsxfun() I couldn't do this...

Sorry I didn't mention about it before... But I need to calculate the distances by using "Euclidean distance" in my function

[sani ars]

thanxx alot Teja Muppirala..... this line is working fine..

out = pdist2(train_data,test_point)....

Answer 2

[Walter Roberson]

Have you considered a "for" loop ?

[sani ars]

No..... In haven't used "for" loop yet...

Do you mean that I should take each row of train_data one by one against test_point??... If yes then, in that case, should I use the function pdist2()??....

[Walter Roberson]

Using one of the library functions to calculate row by row is fine.