Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

DY: Fachverband Dynamik und Statistische Physik

DY 28: Networks: From Topology to Dynamics I (joint session SOE/DY)

DY 28.2: Vortrag

Mittwoch, 20. März 2024, 15:15–15:30, TC 006

ESABO Co-Abundance Analysis: cases where the binarization threshold matters — •Devi Chandran and Jens Christian Claussen — School of Computer Science University of Birmingham, UK

Population dynamics including their complex interactions lead, in societies and microbial populations, to rich co-abundance patterns, and often only the (co)abundance pattern data is measured – whereby the precise interactions remain unknown. Here, ESABO [1] has been introduced to grasp interactions that remain unseen especially for low-abundant species. In [1], using ESABO we have recovered positive and negative interactions between agents (or species) within the population based on co-abundance data. However, in the medium abundance region, instead of a binarization threshold of 1, might it be worth to consider larger binarization thresholds? We investigate, based on two datasets, whether higher thresholds can lead to a higher information gain (in the sense of ESABO), and demonstrate cases of higher information gain for higher thresholds, but also confirm that the original threshold of 1 can be optimal for other datasets.

[1] J.C.Claussen at al., PlosCB (2017) 13(6): e1005361.

Keywords: population dynamics; inferring networks from data; information theory; social networks; microbiome