Machine Learning and Text Mining of Trophic Links

Abstract

Machine Learning has been used to automatically generate a probabilistic food-web from Farm Scale Evaluation (FSE) data. The initial food web proposed by machine learning has been examined by domain experts and comparison with the literature shows that many of the links are corroborated. The FSE data were collected using two different sampling techniques, namely Vortis and pitfall. The corroboration of the initial Vortis food web, generated by machine learning, was performed manually by the domain experts. However, manual corroboration of hypothetical trophic links is difficult and requires significant amounts of time. In this paper we review the method and the main results on machine learning of trophic links. We study common trophic links from Vortis and pitfall data. We also describe a new method and present initial results on automatic corroboration of trophic links using text mining.

Construction and Validation of Food Webs Using Logic-Based Machine Learning and Text Mining

“…Specifically, we will attempt to: (i) develop a robust (general) validation methodology for networks (originally discussed in Tamaddoni-Nezhad et al, 2012b andAfroozi Milani et al, 2012); (ii) grow the network by learning across disparate sampling protocols; and, (iii) develop a generic approach to networks that might allow us to move between systems, based on a more functional, as opposed to a purely taxonomic, approach.…”

Synthetic datasets and community tools for the rapid testing of ecological hypotheses

“…While it is possible to do manual literature surveys (Strong and Leroux 2014), this task becomes daunting for large number of species. Initiatives like text-mining (Milani et al 2012) will speed up the rate at which we can recover interactions data from the literature -if publishers allow researchers to mine the literature they create.…”