SARBayes
19 November 2009, 0:29 UTCTwin Engine Quip
From a list of pilot quotes:
"When one engine fails on a twin-engine airplane, you always have enough power left to get you to the scene of the crash."
read more...
31 May 2009, 9:53 UTCAMDR & ESW II
Updated scatterplots for AMDR vs ESW still show good correlation.
read more...
30 January 2009, 1:44 UTC99% of those found alive...
99% of subjects found alive are found within about 50 hours. Think about that. Should you call off a search after 50 hours?
read more...
5 July 2008, 0:27 UTCAMDR & ESW
23 June 2008, 19:08 UTCProbabilty Mapper README
9 June 2008, 18:10 UTCAutomated Image Segmentation
5 June 2008, 19:47 UTCESRI SAR Articles
8 May 2008, 1:56 UTCMore wireframes
16 February 2008, 11:52 UTCSpreading probability around
18 November 2007, 0:19 UTCInteresting Papers on GIS
2 November 2007, 2:59 UTCSAR Spotlight Forum
22 September 2007, 0:25 UTCRe-Post: Comments on O'Connor
14 September 2007, 13:48 UTCBlog links, MTurk
28 June 2007, 10:52 UTCACFR & Nuit Blanche
25 April 2007, 2:24 UTCRabbit on Search Management
2 April 2007, 20:31 UTCLandmine probability maps
2 April 2007, 10:20 UTCSyrotuck's Data
Corrections and improvements for the
previous wireframes. These have a 40-bin resolution, and fix some errors.
Click the [png | pdf] text to view or download the 300dpi files
Dementia_radial_bivariate_40_bins: [
png |
pdf
]
A 2D normal with the same standard deviation as the Dementia data. Technically, because this is called a bivariate normal distribution it has two variables, x and y. But in SAR literature it is usually called the "circular normal" distribution. Mathematicians tell me that "circular normal" normally means the von Mises distribution, a normal-like distribution for points around a circle. SAR might use this for dispersion angle.
Dementia_radial_pden_40_bins: [
png |
pdf
]
Dementia data, probability density (pDen). Clearly our peak is higher than the matched circular normal (bivariate) distribution. That means we must have more outliers. However, our central peak is so high, they vanish in the scale.
Dementia_radial_num_40_bins: [
png |
pdf
]
Raw proportions: we just rotated the histogram, without correcting for the area of each ring. Also notice how the data is lumpy -- people report data in round numbers, so if you do a cluster analysis, you might want to account for that.
Thanks
Thanks to
njh and
pfh for help.
