Gaussian process convolutions for Bayesian spatial classification
Master's Project (M.S.) University of Alaska Fairbanks, 2016 We compare three models for their ability to perform binary spatial classification. A geospatial data set consisting of observations that are either permafrost or not is used for this comparison. All three use an underlying Gaussian p...
Main Author: | |
---|---|
Other Authors: | , , , |
Format: | Other/Unknown Material |
Language: | English |
Published: |
2016
|
Subjects: | |
Online Access: | http://hdl.handle.net/11122/8031 |
id |
ftunivalaska:oai:scholarworks.alaska.edu:11122/8031 |
---|---|
record_format |
openpolar |
spelling |
ftunivalaska:oai:scholarworks.alaska.edu:11122/8031 2023-05-15T17:57:23+02:00 Gaussian process convolutions for Bayesian spatial classification Best, John K. Short, Margaret Goddard, Scott Barry, Ron McIntyre, Julie 2016-05 http://hdl.handle.net/11122/8031 en_US eng http://hdl.handle.net/11122/8031 Department of Mathematics and Statistics Gaussian processes Spatial analysis (Statistics) Master's Project ms 2016 ftunivalaska 2023-02-23T21:36:58Z Master's Project (M.S.) University of Alaska Fairbanks, 2016 We compare three models for their ability to perform binary spatial classification. A geospatial data set consisting of observations that are either permafrost or not is used for this comparison. All three use an underlying Gaussian process. The first model considers this process to represent the log-odds of a positive classification (i.e. as permafrost). The second model uses a cutoff. Any locations where the process is positive are classified positively, while those that are negative are classified negatively. A probability of misclassification then gives the likelihood. The third model depends on two separate processes. The first represents a positive classification, while the second a negative classification. Of these two, the process with greater value at a location provides the classification. A probability of misclassification is also used to formulate the likelihood for this model. In all three cases, realizations of the underlying Gaussian processes were generated using a process convolution. A grid of knots (whose values were sampled using Markov Chain Monte Carlo) were convolved using an anisotropic Gaussian kernel. All three models provided adequate classifications, but the single and two-process models showed much tighter bounds on the border between the two states. Other/Unknown Material permafrost Alaska University of Alaska: ScholarWorks@UA Fairbanks |
institution |
Open Polar |
collection |
University of Alaska: ScholarWorks@UA |
op_collection_id |
ftunivalaska |
language |
English |
topic |
Gaussian processes Spatial analysis (Statistics) |
spellingShingle |
Gaussian processes Spatial analysis (Statistics) Best, John K. Gaussian process convolutions for Bayesian spatial classification |
topic_facet |
Gaussian processes Spatial analysis (Statistics) |
description |
Master's Project (M.S.) University of Alaska Fairbanks, 2016 We compare three models for their ability to perform binary spatial classification. A geospatial data set consisting of observations that are either permafrost or not is used for this comparison. All three use an underlying Gaussian process. The first model considers this process to represent the log-odds of a positive classification (i.e. as permafrost). The second model uses a cutoff. Any locations where the process is positive are classified positively, while those that are negative are classified negatively. A probability of misclassification then gives the likelihood. The third model depends on two separate processes. The first represents a positive classification, while the second a negative classification. Of these two, the process with greater value at a location provides the classification. A probability of misclassification is also used to formulate the likelihood for this model. In all three cases, realizations of the underlying Gaussian processes were generated using a process convolution. A grid of knots (whose values were sampled using Markov Chain Monte Carlo) were convolved using an anisotropic Gaussian kernel. All three models provided adequate classifications, but the single and two-process models showed much tighter bounds on the border between the two states. |
author2 |
Short, Margaret Goddard, Scott Barry, Ron McIntyre, Julie |
format |
Other/Unknown Material |
author |
Best, John K. |
author_facet |
Best, John K. |
author_sort |
Best, John K. |
title |
Gaussian process convolutions for Bayesian spatial classification |
title_short |
Gaussian process convolutions for Bayesian spatial classification |
title_full |
Gaussian process convolutions for Bayesian spatial classification |
title_fullStr |
Gaussian process convolutions for Bayesian spatial classification |
title_full_unstemmed |
Gaussian process convolutions for Bayesian spatial classification |
title_sort |
gaussian process convolutions for bayesian spatial classification |
publishDate |
2016 |
url |
http://hdl.handle.net/11122/8031 |
geographic |
Fairbanks |
geographic_facet |
Fairbanks |
genre |
permafrost Alaska |
genre_facet |
permafrost Alaska |
op_relation |
http://hdl.handle.net/11122/8031 Department of Mathematics and Statistics |
_version_ |
1766165792518307840 |