Bayesian methods for WiSAR
The "Evaluating Lost Person Behavior" paper is now officially available in the online edition of Transactions in GIS.
The "Evaluating Lost Person Behavior" paper is now officially available in the online edition of Transactions in GIS.
Sava, E., Twardy, C., Koester, R., & Sonwalkar, M. (2016). Evaluating Lost Person Behavior Models. Transactions in GIS, 20(1), 38–53. http://doi.org/10.1111/tgis.12143
The MapScore project described here provides a way to evaluate probability maps using actual historical searches. On a metric where random maps score 0 and perfect maps score 1, the ISRID Distance Ring model scored 0.78 (95%CI: 0.74-0.82, on 376 cases). The Combined model was slightly better at .81 (95%CI: 0.77-0.84).
Our MapScore paper is now in press at Transactions in GIS! From the abstract:
The MapScore project described here provides a way to evaluate probability maps using actual historical searches. In this work we generated probability maps based on the statistical Euclidean distance tables from ISRID data (Koester, 2008), and compared them to Doke’s (2012) watershed model. Watershed boundaries follow high terrain and may better reflect actual barriers to travel. We also created a third model using the joint distribution using Euclidean and watershed features. On a metric where random maps score 0 and perfect maps score 1, the ISRID Distance Ring model scored 0.78 (95%CI: 0.74-0.82, on 376 cases). The simple Watershed model by itself was clearly inferior at 0.61, but the Combined model was slightly better at .81 (95%CI: 0.77-0.84).
Eric Cawi
A logical extension of the Distance Rings model is to fit a smooth function to the distribution of data found in ISRID. Examining the Euclidean Distance data for different categories, it was found that a lognormal curve roughly captured the shape of the data. The Log-Normal (LN) is a two parameter distribution which assumes that the logarithm of your data follows a normal distribution. The probability density function of the LN curve is given by, where are the mean and standard deviation of the logarithm of distance.
People are often incoherent: their probabilities don't add to 100%. We get an 18% gain in accuracy if we coherentize their estimates. But we get a 30% gain in accuracy if we also assign more weight to coherent estimates.
What implications does this have for making subjective probability maps?
The following figure is from a recent paper I co-authored*:
What implications does it have for making subjective "consensus" probability maps at the start of a search?
In which I discuss two-and-a-half approaches to crowdsourcing Search & Rescue, and invite you to try one -- namely, mine: http://bit.ly/1eBBMkq
In which I discuss two-and-a-half approaches to crowdsourcing Search & Rescue, and invite you to try one -- namely, mine. (SciCast)
MapScore has been updated to include Batch Upload, and fix or improve a number of other features.
The MapScore site has been updated! The most exciting new feature is Batch Upload.
Links to Paul Doherty's new research page. Elena Sava has been testing a simple watershed model that Paul helped develop: it scores about 0.55. Considering the simplicity, that's pretty good.
Just a quick note to highlight Paul Doherty's new research page. It includes:
At Mason we're collaborating with Paul to test a Watershed-Distance model developed by his research group. Based on 58 tests run so far by Elena Sava on MapScore, this simple model scores 0.55. Not bad for a model that doesn't yet discriminate by category (or any other feature). Elena just finished a multivariate model combining Watersheds with the more usual crows'-flight distance, and we will begin testing that soon.
The SARBayes MapScore server has been running for a month now at http://mapscore.sarbayes.org. It's a portal for scoring probability maps, so researchers like us can measure how well we are doing, and see which approaches work best for which situations. Take a look. (And if you have a model, register and start testing it!)
Continue reading "MapScore: A Portal for Scoring Probability Maps"
Lin & Goodrich at Brigham Young are working on Bayesian motion models for generating probability maps. They have an interesting model, but need GPS tracks to train it. It's a nice complement to our approach, and it will be interesting to see how they compare.
~Originally a very cool review published in the first half of 2010. The review led to phone calls and a very productive collaboration on MapScore and other work.
Partly reconstructed March 2012.
The following paper creates landmine probability maps almost precisely in the way that I have wanted to create lost person behavior probability maps. (I'll add some more notes on it later.)
