[R-sig-Geo] RE : RE : determining suitable lakes for landing

[Barry Rowlingson]

On Tue, Jul 31, 2012 at 2:21 PM,
Bastien.Ferland-Raymond at mrnf.gouv.qc.ca <Bastien.Ferland-

> The code gave pretty good results, however not perfect.  27 lakes gave different results from the initial Arcinfo code.  Following manually looking at those lakes, I've noticed that 40% were missed by Barry's algorithm (false negative).  However, that also meant that the ArcInfo was doing 60% of false-positive, which makes Barry's code better :)  Increasing the number of simulations to 30000 helped finding some of those.

Considering the application, false positives are a problem! I'd rather
miss out on flying to a lake that I could land on than fly to a lake
that I can't!

 My algorithm shouldn't generate any false positives.

Barry

-- 
blog: http://geospaced.blogspot.com/
web: http://www.maths.lancs.ac.uk/~rowlings
web: http://www.rowlingson.com/
twitter: http://twitter.com/geospacedman
pics: http://www.flickr.com/photos/spacedman