Faster way to calculate the euclidean distance of a matrix than pdist and squareform?

3 次查看（过去 30 天）
Dennis_Pana
Dennis_Pana 2016-7-21
回答： Walter Roberson 2016-7-21
design=[0.625 0.875;0.25 0.75;0.375 0.125;0.875 0.375;0.125 0.625;
0 0;0.5 0.5;0.75 0.25;0.6875 0.0625;0.0625 0.8125;0.1875 0.6875;0.3125 0.3125;
0.4375 0.9375;0.5625 0.5625;0.9375 0.1875;0.8125 0.4375];
v=16;
tic
d(:,:)=squareform(pdist(design(1:v,:))).^2;
toc
This code needs about 0.009 seconds to run. I need something faster, because this calculation is in a for loop. Thank you in advance.

请先登录，再回答此问题。

采纳的回答

Walter Roberson
Walter Roberson 2016-7-21
You can use bsxfun of @minus to do the subtractions, then .^2 to square the components, sum() along the appropriate dimension, and .^(1/2) the result. This should be marginally faster than calling pdist, but probably not much faster.

更多回答（0 个）

类别

在 Help Center 和 File Exchange 中查找有关 Statistics and Machine Learning Toolbox 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by