Honors College Thesis
 

Comparison of machine learning methods for predicting bird distributions

Public Deposited

Downloadable Content

Download PDF
https://ir.library.oregonstate.edu/concern/honors_college_theses/df65v9721

Descriptions

Attribute NameValues
Creator
Abstract
  • The purpose of this study is to explore kernel machine learning methods for species distribution modeling. Previous studies have shown the success of Generalized Boosted Regression Models, however kernel methods have been unexplored for species distribution modeling. Using the eBird dataset, four machine learning methods were tested for accuracy and speed. Accuracy was measured in terms of the Area Under the Curve of the Receiver Operating Characteristic curve. The eBird dataset was divided into training, validation, and testing sets to ensure a fair comparison between the methods, including cross validation for the tuning parameters. The four methods tested were: Generalized Linear Models (GLM), Generalized Boosted Regression Models (GBM), Kernel Support Vector Machines (KSVM), and Kernel Logistic Regression (KLR). The results show that GBM performs better than the kernel methods and the baseline GLM. GBM is not the fastest method, but this is less important since the prediction for species distribution modeling is not a time sensitive matter. Therefore, GBM was found to be the best out of the four methods for species distribution modeling.
License
Resource Type
Date Available
Date Issued
Degree Level
Degree Name
Degree Field
Degree Grantor
Commencement Year
Advisor
Non-Academic Affiliation
Rights Statement
Funding Statement (additional comments about funding)
  • NSF Computational Sustainability Grant
Publisher
Peer Reviewed
Language
Replaces

Relationships

Parents:

This work has no parents.

In Collection:

Items

Thumbnail Title Date Uploaded Visibility Actions
file details FinalCopy.pdf 2017-07-29 Public Download