{"id":{"repo_id":"baylor","oai_identifier":"oai:baylor-ir.tdl.org:2104/15010"},"canonical_url":"https://search.dev.ndltd.org/etd/baylor/oai:baylor-ir.tdl.org:2104/15010","repository":{"repo_id":"baylor","name":"Baylor University","base_url":"https://baylor-ir.tdl.org/server/oai/request"},"display":{"title":"Advances in spatial modeling for complex data with applications to symbolic data and spatial transcriptomics.","abstract":"Spatial statistical modeling is commonly used to analyze and draw inference from data collected across geographic space, providing insight into underlying spatial processes across environmental, biological, epidemiological, and other scientific applications. Spatial data are increasingly complex, high-dimensional, and collected across heterogeneous networks or at varying resolutions, posing challenges for accurate inference, prediction, and biological interpretation that traditional spatial methods are often inadequate to address. This dissertation develops spatial statistical methods that tackle methodological challenges in spatial interval-valued data (SIVD), with particular emphasis on data fusion, integration of multiple data sources and networks, resolution mismatch, and high-dimensional gene-level dependence structures. Specifically, three methodological contributions are proposed and evaluated. First, a harmonized kriging framework for Spatial Interval-Valued Data (SIVD) is proposed to improve spatial prediction by integrating information from heterogeneous monitoring networks while accounting for and correcting systematic inter-network biases. Second, a spatial downscaling framework is developed for SIVD that jointly models the center and range features to address the change-of-support problem through a multivariate smoothing mechanism over the coarse grid, with efficient Bayesian inference facilitated through the Integrated Nested Laplace Approximation (INLA). A bivariate visualization tool is also developed to aid interpretation of the joint behavior of these features across space. Third, SPHERE (Spatial Poisson Hierarchical modEl with pathway-infoRmed gEne networks) is proposed, a Bayesian spatial Poisson lognormal model that jointly captures spatial and gene-level dependencies through pathway-informed Conditional Autoregressive (CAR) priors for the detection of Spatially Expressed (SE) genes in Spatial Transcriptomics data. All proposed models are evaluated through detailed simulation studies under realistic conditions to assess their reliability under known settings, and validated through real data applications to demonstrate their practical applicability. Together, the three frameworks offer efficient and scalable solutions for analyzing complex spatial data, with broad applicability in environmental, public health, and genomics research.","abstract_html":"Spatial statistical modeling is commonly used to analyze and draw inference from data collected across geographic space, providing insight into underlying spatial processes across environmental, biological, epidemiological, and other scientific applications. Spatial data are increasingly complex, high-dimensional, and collected across heterogeneous networks or at varying resolutions, posing challenges for accurate inference, prediction, and biological interpretation that traditional spatial methods are often inadequate to address. This dissertation develops spatial statistical methods that tackle methodological challenges in spatial interval-valued data (SIVD), with particular emphasis on data fusion, integration of multiple data sources and networks, resolution mismatch, and high-dimensional gene-level dependence structures. Specifically, three methodological contributions are proposed and evaluated. First, a harmonized kriging framework for Spatial Interval-Valued Data (SIVD) is proposed to improve spatial prediction by integrating information from heterogeneous monitoring networks while accounting for and correcting systematic inter-network biases. Second, a spatial downscaling framework is developed for SIVD that jointly models the center and range features to address the change-of-support problem through a multivariate smoothing mechanism over the coarse grid, with efficient Bayesian inference facilitated through the Integrated Nested Laplace Approximation (INLA). A bivariate visualization tool is also developed to aid interpretation of the joint behavior of these features across space. Third, SPHERE (Spatial Poisson Hierarchical modEl with pathway-infoRmed gEne networks) is proposed, a Bayesian spatial Poisson lognormal model that jointly captures spatial and gene-level dependencies through pathway-informed Conditional Autoregressive (CAR) priors for the detection of Spatially Expressed (SE) genes in Spatial Transcriptomics data. All proposed models are evaluated through detailed simulation studies under realistic conditions to assess their reliability under known settings, and validated through real data applications to demonstrate their practical applicability. Together, the three frameworks offer efficient and scalable solutions for analyzing complex spatial data, with broad applicability in environmental, public health, and genomics research.","abstract_has_math":false,"creators":["Sarfo Fosu, Emmanuel, 1994-"],"institution":"Baylor University.","degree_name":"Ph.D.","degree_level":"Doctoral","degree_discipline":null,"degree_department":null,"school":null,"contributors":[],"advisors":["Song, Joon Jin."],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026-08","date_published":"2026-08","updated_at":"2026-07-24T01:07:54Z","subjects":["Spatial statistics.","Data harmonization.","Symbolic data.","Spatial transcriptomics.","Downscaling."],"languages":["en"],"rights":[],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/2104/15010","outbound_label":"Handle","outbound_source":"dc:identifier.uri"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Song, Joon Jin."]},{"key":"dc:creator","label":"Author","values":["Sarfo Fosu, Emmanuel, 1994-"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-06-14T00:54:42Z"]},{"key":"dc:date.issued","label":"Date","values":["2026-08"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Doctoral"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["Baylor University."]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Spatial statistics.","Data harmonization.","Symbolic data.","Spatial transcriptomics.","Downscaling."]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language.iso","label":"Language (ISO)","values":["en"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/2104/15010"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Spatial statistical modeling is commonly used to analyze and draw inference from data collected across geographic space, providing insight into underlying spatial processes across environmental, biological, epidemiological, and other scientific applications. Spatial data are increasingly complex, high-dimensional, and collected across heterogeneous networks or at varying resolutions, posing challenges for accurate inference, prediction, and biological interpretation that traditional spatial methods are often inadequate to address. This dissertation develops spatial statistical methods that tackle methodological challenges in spatial interval-valued data (SIVD), with particular emphasis on data fusion, integration of multiple data sources and networks, resolution mismatch, and high-dimensional gene-level dependence structures. Specifically, three methodological contributions are proposed and evaluated. First, a harmonized kriging framework for Spatial Interval-Valued Data (SIVD) is proposed to improve spatial prediction by integrating information from heterogeneous monitoring networks while accounting for and correcting systematic inter-network biases. Second, a spatial downscaling framework is developed for SIVD that jointly models the center and range features to address the change-of-support problem through a multivariate smoothing mechanism over the coarse grid, with efficient Bayesian inference facilitated through the Integrated Nested Laplace Approximation (INLA). A bivariate visualization tool is also developed to aid interpretation of the joint behavior of these features across space. Third, SPHERE (Spatial Poisson Hierarchical modEl with pathway-infoRmed gEne networks) is proposed, a Bayesian spatial Poisson lognormal model that jointly captures spatial and gene-level dependencies through pathway-informed Conditional Autoregressive (CAR) priors for the detection of Spatially Expressed (SE) genes in Spatial Transcriptomics data. All proposed models are evaluated through detailed simulation studies under realistic conditions to assess their reliability under known settings, and validated through real data applications to demonstrate their practical applicability. Together, the three frameworks offer efficient and scalable solutions for analyzing complex spatial data, with broad applicability in environmental, public health, and genomics research."]},{"key":"dc:format.mimetype","label":"Dc Format Mimetype","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Advances in spatial modeling for complex data with applications to symbolic data and spatial transcriptomics."]}]}],"canonical_facts":{"dc:contributor.advisor":["Song, Joon Jin."],"dc:creator":["Sarfo Fosu, Emmanuel, 1994-"],"dc:date.accessioned":["2026-06-14T00:54:42Z"],"dc:date.issued":["2026-08"],"dc:description.abstract":["Spatial statistical modeling is commonly used to analyze and draw inference from data collected across geographic space, providing insight into underlying spatial processes across environmental, biological, epidemiological, and other scientific applications. Spatial data are increasingly complex, high-dimensional, and collected across heterogeneous networks or at varying resolutions, posing challenges for accurate inference, prediction, and biological interpretation that traditional spatial methods are often inadequate to address. This dissertation develops spatial statistical methods that tackle methodological challenges in spatial interval-valued data (SIVD), with particular emphasis on data fusion, integration of multiple data sources and networks, resolution mismatch, and high-dimensional gene-level dependence structures. Specifically, three methodological contributions are proposed and evaluated. First, a harmonized kriging framework for Spatial Interval-Valued Data (SIVD) is proposed to improve spatial prediction by integrating information from heterogeneous monitoring networks while accounting for and correcting systematic inter-network biases. Second, a spatial downscaling framework is developed for SIVD that jointly models the center and range features to address the change-of-support problem through a multivariate smoothing mechanism over the coarse grid, with efficient Bayesian inference facilitated through the Integrated Nested Laplace Approximation (INLA). A bivariate visualization tool is also developed to aid interpretation of the joint behavior of these features across space. Third, SPHERE (Spatial Poisson Hierarchical modEl with pathway-infoRmed gEne networks) is proposed, a Bayesian spatial Poisson lognormal model that jointly captures spatial and gene-level dependencies through pathway-informed Conditional Autoregressive (CAR) priors for the detection of Spatially Expressed (SE) genes in Spatial Transcriptomics data. All proposed models are evaluated through detailed simulation studies under realistic conditions to assess their reliability under known settings, and validated through real data applications to demonstrate their practical applicability. Together, the three frameworks offer efficient and scalable solutions for analyzing complex spatial data, with broad applicability in environmental, public health, and genomics research."],"dc:format.mimetype":["application/pdf"],"dc:identifier.uri":["https://hdl.handle.net/2104/15010"],"dc:language.iso":["en"],"dc:subject":["Spatial statistics.","Data harmonization.","Symbolic data.","Spatial transcriptomics.","Downscaling."],"dc:title":["Advances in spatial modeling for complex data with applications to symbolic data and spatial transcriptomics."],"dc:type":["Thesis"],"thesis:degree_level":["Doctoral"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["Baylor University."]},"updated_at":"2026-07-24T01:07:54Z"}