-
Created
September 05, 2021 17:20
by
axm733
-
Added Creator Arif Masrur
September 05, 2021 17:21
by
axm733
-
Added
manuscript_non-anonymous.pdf
September 05, 2021 17:24
by
axm733
-
Added
Supplementary information-iST-RF.pdf
September 05, 2021 17:24
by
axm733
-
September 05, 2021 17:29
by
axm733
Publication Date
License
- https://opensource.org/licenses/MIT
-
Added Creator Manzhu Yu
September 05, 2021 17:42
by
axm733
-
September 07, 2021 20:47
by
axm733
Keyword
- Spatial modeling, Machine learning interpretability , Spatial heterogeneity, Random forest, Decision tree, Wildfire
Geographic Area
United States
- Continental U.S.
-
September 07, 2021 20:49
by
axm733
Acknowledgments
- We acknowledge Penn State Center for Security Research and Education (CSRE) and Information Sciences and Technology (IST) for providing seed grants to conduct this study.
-
September 07, 2021 20:55
by
axm733
Description
ABSTRACT
- This preprint version precedes the formal peer review and publication in IJGIS. Final version: https://doi.org/10.1080/13658816.2021.1965608
- ABSTRACT
- Machine learning (ML) interpretability has become increasingly crucial for identifying accurate and relevant structural relationships between spatial events and factors that explain them. Methodologically aspatial ML algorithms with an apparent high predictive power ignore characteristic non-stationary domain relationships in spatio-temporal data (e.g., dependence, heterogeneity), and can lead to incorrect interpretations and poor management decisions. This study addresses this critical methodological issue of ‘interpretability’ in ML-based modeling of structural relationships between spatio-temporal events and corresponding bio-physical drivers. Specifically, we present and evaluate a spatio-temporally interpretable random forest (iST-RF) modeling framework using the example of drivers of wildfire characteristics across the United States. Our experiments show that the spatio-temporal sampling and weighted prediction approach can improve predictive accuracy (76%) compared to the aspatial RF approach (70%), while also enhancing interpretations of the ML model’s spatio-temporal relevance for its ensemble prediction. This novel approach can help balance prediction and interpretation with fidelity in a spatial data science life cycle. However, challenges exist for predictive modeling when dataset is very small, because in such cases locally optimized sub-model’s prediction performance can be suboptimal. With that caveat, our proposed approach is an ideal choice for identifying drivers of spatio-temporal events at country or regional-scale studies.
-
September 07, 2021 20:56
by
axm733
License
https://opensource.org/licenses/MIT
- https://creativecommons.org/licenses/by-nc-nd/4.0/
-
September 07, 2021 20:57
by
axm733
Description
- This preprint version precedes the formal peer review and publication in IJGIS. Final version: https://doi.org/10.1080/13658816.2021.1965608
ABSTRACT
Machine learning (ML) interpretability has become increasingly crucial for identifying accurate and relevant structural relationships between spatial events and factors that explain them. Methodologically aspatial ML algorithms with an apparent high predictive power ignore characteristic non-stationary domain relationships in spatio-temporal data (e.g., dependence, heterogeneity), and can lead to incorrect interpretations and poor management decisions. This study addresses this critical methodological issue of ‘interpretability’ in ML-based modeling of structural relationships between spatio-temporal events and corresponding bio-physical drivers. Specifically, we present and evaluate a spatio-temporally interpretable random forest (iST-RF) modeling framework using the example of drivers of wildfire characteristics across the United States. Our experiments show that the spatio-temporal sampling and weighted prediction approach can improve predictive accuracy (76%) compared to the aspatial RF approach (70%), while also enhancing interpretations of the ML model’s spatio-temporal relevance for its ensemble prediction. This novel approach can help balance prediction and interpretation with fidelity in a spatial data science life cycle. However, challenges exist for predictive modeling when dataset is very small, because in such cases locally optimized sub-model’s prediction performance can be suboptimal. With that caveat, our proposed approach is an ideal choice for identifying drivers of spatio-temporal events at country or regional-scale studies.
- ABSTRACT: Machine learning (ML) interpretability has become increasingly crucial for identifying accurate and relevant structural relationships between spatial events and factors that explain them. Methodologically aspatial ML algorithms with an apparent high predictive power ignore characteristic non-stationary domain relationships in spatio-temporal data (e.g., dependence, heterogeneity), and can lead to incorrect interpretations and poor management decisions. This study addresses this critical methodological issue of ‘interpretability’ in ML-based modeling of structural relationships between spatio-temporal events and corresponding bio-physical drivers. Specifically, we present and evaluate a spatio-temporally interpretable random forest (iST-RF) modeling framework using the example of drivers of wildfire characteristics across the United States. Our experiments show that the spatio-temporal sampling and weighted prediction approach can improve predictive accuracy (76%) compared to the aspatial RF approach (70%), while also enhancing interpretations of the ML model’s spatio-temporal relevance for its ensemble prediction. This novel approach can help balance prediction and interpretation with fidelity in a spatial data science life cycle. However, challenges exist for predictive modeling when dataset is very small, because in such cases locally optimized sub-model’s prediction performance can be suboptimal. With that caveat, our proposed approach is an ideal choice for identifying drivers of spatio-temporal events at country or regional-scale studies.
-
September 07, 2021 20:59
by
axm733
-
Published
September 07, 2021 21:00
by
axm733
-
Added Creator Prasenjit Mitra
November 11, 2021 09:20
by
sre53
-
Added Creator Donna Jean Peuquet
November 11, 2021 09:20
by
sre53
-
Added Creator Alan H Taylor
November 11, 2021 09:20
by
sre53
-
Updated
March 22, 2022 16:18
by
[unknown user]
-
Updated
April 04, 2024 10:21
by
[unknown user]