Hopeful Colloborative Project Upcoming With AU

I started working with a professor at AU today on trying to spatially predict some measurements by krigging.  The problem I was asked to address involved two different sized measurements and relating these together.  I think the researchers were wanting to predict a small area by measuring a large area and vis versa.  I suggested using possibly three different techniques to produce a common krigging surface.  The most popular of these three tools is called Delauney Triangles.  This takes three measurements and produces a new measurement weighted by the distance from each point and relabels this measurement point in the center of the triangle.  The other two methods were Thiessen polygons and cell declusterings.  The problem with Thiessen polygons falls in that the krigging surface will become discontinous which doesn’t look smooth.  This will also produce terrible estimates at discontiunity portions of the krigging surface.  Cell declustering, dividing the entire prediction region into equally spaced regions, has a problem because the desclustering window size is purely up to the researcher.  There is no purely objective way to determine declustering window size and in essence, the researher can produce any size variant measurements they want.  This variance has strong statisitcal implications.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s