{"id":{"repo_id":"buffalo","oai_identifier":"oai:ubir.buffalo.edu:10477/79335"},"canonical_url":"https://search.dev.ndltd.org/etd/buffalo/oai:ubir.buffalo.edu:10477/79335","repository":{"repo_id":"buffalo","name":"Buffalo","base_url":"https://ubir.buffalo.edu/oai/request"},"display":{"title":"Approaches to Extract Deep Phenotypes from Clinical Data","abstract":"Ph.D.","abstract_html":"Ph.D.","abstract_has_math":false,"creators":["Luong, Duc Thanh Anh; 0000-0003-4768-5089"],"institution":"State University of New York at Buffalo","degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Chandola, Varun","Computer Science and Engineering"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2019,"date_issued":"2019-04-04T20:30:28Z","date_published":"2019-04-04T20:30:28Z","updated_at":"2026-07-27T19:05:14Z","subjects":["computer science","biomedical engineering","computer engineering"],"languages":["eng"],"rights":["Users of works found in University at Buffalo Institutional Repository (UBIR) are responsible for identifying and contacting the copyright owner for permission to reuse. University at Buffalo Libraries do not manage rights for copyright-protected works and cannot assist with permissions.","Copyright retained by author."],"rights_urls":[],"identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/10477/79335","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Chandola, Varun","Computer Science and Engineering"]},{"key":"dc:creator","label":"Author","values":["Luong, Duc Thanh Anh; 0000-0003-4768-5089"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2019-04-04T20:30:28Z","2019","2018-10-14 23:22:29"]},{"key":"dc:publisher","label":"Institution","values":["State University of New York at Buffalo"]},{"key":"dc:type","label":"Dc Type","values":["Text","Dissertation"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["computer science","biomedical engineering","computer engineering"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]},{"key":"dc:rights","label":"Dc Rights","values":["Users of works found in University at Buffalo Institutional Repository (UBIR) are responsible for identifying and contacting the copyright owner for permission to reuse. University at Buffalo Libraries do not manage rights for copyright-protected works and cannot assist with permissions.","Copyright retained by author."]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/10477/79335"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["Ph.D.","Computational phenotyping is an emerging topic in health informatics. An important catalyst of this emergence is the increasing volume of clinical data available for analysis. However, clinical data typically consists of many disparate elements and has strong temporal dependencies, both of which require designing new machine learning algorithms that can extract useful knowledge from such data. In this dissertation, my focus is on developing techniques to extract longitudinal phenotypes from clinical data. In particular, we have examined three different approaches to obtain longitudinal disease subtypes for Chronic Kidney Disease (CKD) - a rising health problem in both US and worldwide. First, a probabilistic model is applied to empirically obtained phenotypes by computing the disease subtyping effect while \"explaining away\" other factors such as subtyping effect, long-term and short-term effects from clinical observations.Second, a temporal k-means algorithm was proposed to group patients with similar disease progressions into clusters. Third, we perform an in-depth analysis of Time-Aware Long Short-Term Memory Autoencoder, a deep learning approach, to project longitudinal patient profiles into a common latent space and subsequently use latent representations to identify unusual disease progressions. In order to evaluate the quality of disease subtypes obtained from these approaches, we also propose a quantitative evaluation metric by estimating the tightness of the resulting clusters as well as the degree of separation between different clusters."]},{"key":"dc:format","label":"Dc Format","values":["application/pdf"]},{"key":"dc:title","label":"Title","values":["Approaches to Extract Deep Phenotypes from Clinical Data"]}]}],"canonical_facts":{"dc:contributor":["Chandola, Varun","Computer Science and Engineering"],"dc:creator":["Luong, Duc Thanh Anh; 0000-0003-4768-5089"],"dc:date":["2019-04-04T20:30:28Z","2019","2018-10-14 23:22:29"],"dc:description":["Ph.D.","Computational phenotyping is an emerging topic in health informatics. An important catalyst of this emergence is the increasing volume of clinical data available for analysis. However, clinical data typically consists of many disparate elements and has strong temporal dependencies, both of which require designing new machine learning algorithms that can extract useful knowledge from such data. In this dissertation, my focus is on developing techniques to extract longitudinal phenotypes from clinical data. In particular, we have examined three different approaches to obtain longitudinal disease subtypes for Chronic Kidney Disease (CKD) - a rising health problem in both US and worldwide. First, a probabilistic model is applied to empirically obtained phenotypes by computing the disease subtyping effect while \"explaining away\" other factors such as subtyping effect, long-term and short-term effects from clinical observations.Second, a temporal k-means algorithm was proposed to group patients with similar disease progressions into clusters. Third, we perform an in-depth analysis of Time-Aware Long Short-Term Memory Autoencoder, a deep learning approach, to project longitudinal patient profiles into a common latent space and subsequently use latent representations to identify unusual disease progressions. In order to evaluate the quality of disease subtypes obtained from these approaches, we also propose a quantitative evaluation metric by estimating the tightness of the resulting clusters as well as the degree of separation between different clusters."],"dc:format":["application/pdf"],"dc:identifier":["http://hdl.handle.net/10477/79335"],"dc:language":["eng"],"dc:publisher":["State University of New York at Buffalo"],"dc:rights":["Users of works found in University at Buffalo Institutional Repository (UBIR) are responsible for identifying and contacting the copyright owner for permission to reuse. University at Buffalo Libraries do not manage rights for copyright-protected works and cannot assist with permissions.","Copyright retained by author."],"dc:subject":["computer science","biomedical engineering","computer engineering"],"dc:title":["Approaches to Extract Deep Phenotypes from Clinical Data"],"dc:type":["Text","Dissertation"]},"updated_at":"2026-07-27T19:05:14Z"}