Building and validating bayesian models
Download
1 / 14

Building and Validating Bayesian Models - PowerPoint PPT Presentation


  • 83 Views
  • Uploaded on

Building and Validating Bayesian Models. Identification of Mountain Goat Winter Range in North-central BC. Acknowledgements. Funding from the BC Min. of Environment Other participants included: R. Ellis

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about ' Building and Validating Bayesian Models' - emmanuel-hansen


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
Building and validating bayesian models

Building and Validating Bayesian Models

Identification of Mountain Goat Winter Range in North-central BC

R. Scott McNay, Wildlife Infometrics

Randy Sulyma, BC Min. of Forests


Acknowledgements
Acknowledgements

  • Funding from the BC Min. of Environment

  • Other participants included:

    • R. Ellis

    • D. Fillier, S. Gordon, L. Vanderstar, D. Heard, G. Watts, D. Wilson, J. Vinnedge, B. Brade, R. MacDonald

    • Line Giguere, Robin McKinley

  • Concepts and ideas:

    • the last workshop in Chase

    • subsequent discussions, most notably, B. Marcot & S. Wilson, C. Apps


Modeling context

Theory

Correlative

Theory

Mechanistic

Frequency

Probability

Empirical

Mechanistic

Empirical

Correlative

Modeling Context

Conceptual

from Bunnell 1989

  • Uses:

    • Prediction

      • Management

      • Implications of Predictions?

    • Explanation

      • Research

      • Why?


Rationale
Rationale

  • General, portable model

  • Management & research

  • Prediction & explanation

  • Minimal resources to develop

    • Little information from FSJ

    • Insufficient resources to develop empirical or other more traditional approaches

    • Limited Time Frame


Study areas
Study Areas

  • Adjacent MUs

  • Similar but not the same

  • Preliminary model already built


Simple uwr model
Simple UWR Model

= Spatial relationships of cells

processed/defined in a GIS.

Typically a distance function

from escape terrain.


= Spatial relationships of cells

processed/defined in a GIS



Model construction results
Model Construction Results

Primary indicators:

  • Accuracy

    • 100% relocations

    • Maximize coverage

  • Precision

    • 100% alpine

    • Maximize area decrease


Overall results
Overall Results:

  • Basic model of EP

    • 65% relocations covered

    • 57% reduction in alpine

  • Spatially generalized model (nearest neighbor algorithm)

    • 92% relocations covered

    • 70% reduction in alpine


Model testing
Model Testing

  • Random sample approach applied.

  • Aerial reconnaissance completed.

  • Data collected to evaluate/verify both input parameters, and summary results.

  • Given funding and timing constraints, not possible to evaluate some of the spatial relationships.


Model testing results
Model Testing Results

Modeled

Correct Classification Rate = 67%

False Positive Error Rate = 73%

False Negative Error Rate = 0%

Κ = 0.50

τ = 0.32


Model testing results1
Model Testing Results

Observed

Modeled

Correct Classification Rate = 89%

False Positive Error Rate = 46%

False Negative Error Rate = 0%

Κ = 0.64

τ = 0.79


Discussion
Discussion

  • Were we able to restrict our search for UWR sufficiently yet remain accurate?

  • Was it important that we were not strictly analytical in our Bayesian learning?

  • Was our test protocol appropriate given the project goal?


ad