Biology subjects

Hohmann, N.

Publications and source records attributed to Hohmann, N..

3 recordsLinked to original sources

bioRxiv2018

Reversing Time Averaging and Reconstructing Extinction Rates with Approaches from Image Processing

In this paper, the relation between the extinction rate and the rate of last fossil occurrences as well as the relation between the fossil occurrence rate and the time averaged fossil occurrence rate is examined. Both relations are described by the same mathematical operation. This operation is commonly used in image processing, where it generates a blurring effect. Therefore the rate of last fossil occurrences can be taken as a blurred version of the extinction rate, and the time averaged fossil occurrence rate as a blurred version of the fossil occurrence rate. This connection has different applications. It allows to study the patterns different types of time averaging generate or the patterns of last fossil occurrences generated by different extinction rates. More importantly, it opens the possibility to use algorithms from image processing that reverse blurring effects for geological applications. This can be used to reverse the effects of time averaging or to reconstruct extinction rates from the rate of last fossil occurrences.

paleontology↗

bioRxiv2018

Enforced Symmetry: The Necessity of Symmetric Waxing and Waning

This paper demonstrates that the symmetric waxing and waning pat-terns that can be observed in ecological measures such as occupancy and geographic range are created by averaging, rescaling, conditioning and combining different sources of information. Therefore symmetric waxing and waning from the origination of a taxon to the extinction of a taxon in any measure should be treated as the null hypothesis and non-informative.

ecology↗

bioRxiv2017

Testing hypotheses on population dynamics: A test for the inhomogeneous Poisson point process model

In this paper, a test for hypotheses on population dynamics is presented alongside an implementation of said test for R. The test is based on the assumption that the sample, consisting of points on a time axis, is a realization of a Poisson point process (PPP). There are no restrictions on the shapes of the rate functions that are regulating the PPP, type 2 errors can be calculated and the test is optimal in the sense that it is a uniform most powerful (UMP) test. So for every significance level a, the presented test has a lower type 2 error than every other test having the same significance level a. The test is applicable to all models based on PPPs, including models in spatial dimensions. It can be generalized and expanded in different ways, such as testing larger hypotheses, incorporating prior knowledge, and constructing confidence regions that can be used to obtain upper or lower bounds on rate functions.

bioinformatics↗