It seems that I had a serious bias towards Matlab and I didn’t got the real spirit of it. Now I need to learn the right way to use Matlab.
I want to share some tricks for making Matlab function more efficient and robust. I decide to write a series of blog posts. This is the first one of this series, in which I want to show a simple function for computing pairwise Euclidean distances between points in high dimensional vector space.
The most important thing for efficient Matlab code is VECTORIZATION! Here I will take distance function as an example to demonstrate how accelerate your code by getting ride of “for-loops”.
Say, we have two sets of points X and Y in d-dimensional space. We want to compute the distances between any point in X and any other point in Y. This function is very useful. For example, if you want to find the k-nearest-neighbor of a point x in a database Y, an naive method is to first compute the distances between the point x and all points in Y…
View original post 335 more words
Simplex Signum Veri 