Quick start for deheaping/ Introduction continued

With people from the Illinois Natural History Survey becoming involved in the deheaping project, dealing with data with “spikes” related to bag or other limits became an issue. These spikes are not heaps due to prototype use. To address prototype deheaping in the presence of such spikes, a version of the deheaping program has been developed in which both spikes, irregular protrusion values, and prototype heaps can be specified. If spiking and heaping are occurring at multiples of 5 or 7, the program requires specification of the relative size of heap compared to spike. You guess from the pattern you see and then can adjust ratios in several runs (automation is not a priority).

The Illinois data also introduced the matter of dealing with data with large values and lots of zero frequencies for large values. Because of “scuba” data, a version of the deheaping program using heavy tail estimation (our term) is in development. In fact, the November 2015 deheaping progarm version has a built in a module for dealing with “excessive” zeros.  Having to deal with excessive zeros resulted in a lot of insights. Some are commented on in the “User guide/Documentation” for the November 2015 Deheaping program.

When/if you move on to access a deheaping program, the “current best” is still listed/shown as the program with limited features. However, for ease of running deheaping the alternative program is better both in program and documentation. Runs to this point have eliminated many (hopefully most) problems with using the program. Still, the program is considered “Beta.”

The development team strongly recommends downloading the ZIP with the November 2015 version of the deheaping program. The ZIP has documentation on use. One learns about the interface and issues in deheaping using example data. etc. The learning material has by 3 step by step examples you run using data and guidance supplied. There is new material dealing with theory using simulation results,  some theory, etc. Again, you can preview by downloading (previous page).