Graduate Project
 

Applications of Conditional Topic Models to Species Distribution Prediction

Öffentlich Deposited

Herunterladbarer Inhalt

PDF Herunterladen
https://ir.library.oregonstate.edu/concern/graduate_projects/79407x24n

Descriptions

Attribute NameValues
Creator
Abstract
  • The topic of species distribution modelling has been on of increasing interest in recent years. As climate change is becoming of even more interest to researchers, more tools are needed to better analyze and predict various climate change scenarios. One particular area of interest is that of species distribution modeling. Species distribution modelling addresses the problem of determining either the fundamental or the realized niche of a species, either at the current time or projecting into the past or future. Species distribution models (SDMs) are seen as a potentially powerful tool both for applied policy decisions like reservation design and theorectical understanding, discovering what factors are most important in determining the fundamental niche of a species, as well as the extent to which various factors determine how much of that niche is realized. Currently, almost all SDMs focus on a single species at a time. For any given species, a model is developed and trained for that particular species. An advantage of this approach is that is keep computational costs down relative to a broader model. There is, however, potential in the idea that by modeling multiple species at once, mutual information between species can be leveraged to provide more accurate modeling while offering insights into the nature of the relationships between specific species. This paper examines the attempt to use one such model for doing species distribution modeling on several species at once.
Resource Type
Date Available
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Academic Affiliation
Urheberrechts-Erklärung
Publisher
Peer Reviewed
Language
Replaces

Beziehungen

Parents:

This work has no parents.

Artikel

Miniaturansicht Titel Hochladedatum Sichtbarkeit Aktionen
file details wilkins.11232010.project-2.pdf 2017-07-05 Öffentlich Herunterladen