Rank: 3193 based on 33 downloads (last 30 days) and 2 files submitted
photo

FDR

E-mail
Company/University
EPFL

Personal Profile:
Professional Interests:
optimization

 

Watch this Author's files

 

Files Posted by FDR View all
Updated   File Tags Downloads
(last 30 days)
Comments Rating
25 Nov 2011 Screenshot Local Neighborhood Filters data set to only include local points around a reference. Author: FDR locality filter, data processing 11 2
18 Nov 2011 Screenshot Data Convexity Check Checks if given data may be represented by a convex function within a specified tolerance. Author: FDR convex interpolation, convexity verificatio..., convex fitting 22 0
Comments and Ratings by FDR
Updated File Comments Rating
22 Nov 2011 Local Neighborhood Filters data set to only include local points around a reference. Author: FDR

Hey Sven,

You're right. I tried yours for a 5000x5000 example with both 2 and 3 norms, and it seems to consistently perform about 20-30% faster.

However, you've now inspired me to try to improve on my original a little more (using upper and lower bounds in the looping procedure to immediately remove large parts of the data without analyzing it all). It should be faster, if it works. If not, then I'll update my file with your suggestion.

Comments and Ratings on FDR's Files View all
Updated File Comment by Comments Rating
22 Nov 2011 Local Neighborhood Filters data set to only include local points around a reference. Author: FDR FDR

Hey Sven,

You're right. I tried yours for a 5000x5000 example with both 2 and 3 norms, and it seems to consistently perform about 20-30% faster.

However, you've now inspired me to try to improve on my original a little more (using upper and lower bounds in the looping procedure to immediately remove large parts of the data without analyzing it all). It should be faster, if it works. If not, then I'll update my file with your suggestion.

21 Nov 2011 Local Neighborhood Filters data set to only include local points around a reference. Author: FDR Sven

Good work with the clear help, but note that your loop can be replaced with the (relatively faster):
distsSq = sum(bsxfun(@minus, U, u0).^2,2);
wheat = find(distsSq <= lim^2);
[~,order] = sort(distsSq(wheat));
Usort = U(wheat(order),:);

Contact us