Spatial neural network: Difference between revisions

Revision as of 03:37, 4 May 2024

Latest revision as of 23:02, 17 June 2025

(2 intermediate revisions by 2 users not shown)

Line 3:

{{other uses|SNN (disambiguation)}}

[[File:GWNN and GWR prediction differences.jpg |thumb |upright=1.50 |Difference in predicted house prices within the states of Austria, from a GWR and a GWNN whose the weighting metrics respectively use the Euclidean distance (ED) and travel time distance (TTD)<ref name="Hagenauer et al. (2022)" />]]

<!-- please be cautious in revising the lead/introduction since its visible and invisible texts transclude in the entry/article on: 'spatial analysis' and 'types of artificial neural networks'; please read the webpages: 'Wikipedia:INCLUDEONLY' and 'Wikipedia:PARTRANS', for understanding the properties and purposes of the used HTML tags --><onlyinclude><noinclude>'''Spatial neural networks''' ('''SNNs''')</noinclude><includeonly>Spatial neural networks (SNNs)</includeonly> constitute a supercategory of tailored [[artificial neural networks|neural networks (NNs)]] for representing and predicting geographic phenomena. They generally improve both the statistical [[Accuracy and precision|accuracy]] and [[Statistical reliability|reliability]] of the a-spatial/classic NNs whenever they handle [[Geographic data and information| geo-spatial datasets]]<!-- if you transform 'geo-spatial' into 'geospatial' or conversely, please apply the transformation everywhere -->, and also of the other spatial [[Statistical model|(statistical) models]] (e.g. spatial regression models) whenever the geo-spatial [[data set|datasets]]' variables depict [[Nonlinear system|non-linear relations]].<ref name="Morer et al. (2020)">{{cite journal |vauthors=Morer I, Cardillo A, Díaz-Guilera A, Prignano L, Lozano S |date=2020 |title=Comparing spatial networks: a one-size-fits-all efficiency-driven approach |journal=Physical Review |volume=101 |issue=4 |page=042301 |doi=10.1103/PhysRevE.101.042301|pmid=32422764 |bibcode=2020PhRvE.101d2301M |hdl=2445/161417 |s2cid=49564277 |hdl-access=free |arxiv=1807.00565 }}</ref><ref name="Gupta et al. (2021)">{{cite journal |vauthors=Gupta J, Molnar C, Xie Y, Knight J, Shekhar S |date=2021 |title=Spatial variability aware deep neural networks (SVANN): a general approach |journal=ACM Transactions on Intelligent Systems and Technology |volume=12 |issue=6 |pages=1&ndash;21 |doi=10.1145/3466688|s2cid=244786699 }}</ref><ref name="Hagenauer et al. (2022)">{{cite journal |vauthors=Hagenauer J, Helbich M |date=2022 |title=A geographically weighted artificial neural network |journal=International Journal of Geographical Information Science |volume=36 |issue=2 |pages=215&ndash;235 |doi=10.1080/13658816.2021.1871618|s2cid=233883395 |doi-access=free |bibcode=2022IJGIS..36..215H }}</ref><includeonly> Examples of SNNs are the OSFA spatial neural networks, SVANNs and GWNNs.</includeonly></onlyinclude>

==History==

Line 14:

There exist several categories of methods/approaches for designing and applying SNNs.

*'''One-Size-Fits-all (OSFA) spatial neural networks''', use the OSFA method/approach for globally computing the spatial [[Weighting|weights]] and designing a spatial [[structure]] from the originally a-spatial/classic neural networks.<ref name="Morer et al. (2020)"/>

*'''Spatial Variability Aware Neural Networks''' ('''SVANNs''') use an enhanced OSFA method/approach that locally recomputes the spatial weights and redesigns the spatial structure of the originally a-spatial/classic NNs, at each geo-___location of the (statistical) individuals/units' attributes' values.<ref name="Gupta et al. (2021)"/> They generally outperform the OSFA spatial neural networks, but they do not consistently handle the spatial heterogeneity at multiple scales.<ref name="Xie et al. (2023)">{{cite journal |vauthors=Xie Y, Chen W, He E, Jia X, Bao H, Zhou X, Ghosh E, Ravirathinam P |date=2023 |title=Harnessing heterogeneity in space with statistically guided meta-learning |journal=Knowledge and Information Systems |volume=65 |issue=6 |pages=2699–2729 |doi=10.1007/s10115-023-01847-0|pmid=37035130 |s2cid=257436979 |pmc=9994417 |bibcode=2023KIS....65.2699X }}</ref>

*'''Geographically Weighted Neural Networks''' ('''GWNNs''') are similar to the SVANNs but they use the so-called Geographically Weighted Model (GWM) method/approach by Lu et al. (2023), so to locally recompute the spatial weights and redesign the spatial structure of the originally a-spatial/classic neural networks.<ref name="Hagenauer et al. (2022)"/><ref name="Lu et al. (2023)">{{cite journal |vauthors=Lu B, Hu Y, Yang D, Liu Y, Liao L, Yin Z, Xia T, Dong Z, Harris P, Brunsdon C, Comber A, Dong G |date=2023 |title=GWmodelS: A software for geographically weighted models |journal=SoftwareX |volume=21 |page=101291 |doi=10.1016/j.softx.2022.101291|bibcode=2023SoftX..2101291L |url=https://eprints.whiterose.ac.uk/194864/7/1-s2.0-S2352711022002096-main.pdf }}</ref> Like the SVANNs, they do not consistently handle spatial heterogeneity at multiple scales.<ref name="Hagenauer et al. (2022)"/>

Line 38:

{{reflist}}

[[Category:Artificial neural networks]]

[[Category:Neural network architectures]]

[[Category:Spatial analysis]]