{"id":{"repo_id":"mcmaster","oai_identifier":"oai:macsphere.mcmaster.ca:11375/30264"},"canonical_url":"https://search.ndltd.org/etd/mcmaster/oai:macsphere.mcmaster.ca:11375/30264","repository":{"repo_id":"mcmaster","name":"McMaster University","base_url":"https://macsphere.mcmaster.ca/server/oai/request"},"display":{"title":"Discovery of Temporal Graph Functional Dependencies","abstract":"Temporal graphs play a crucial role in modelling dynamic applications where (node) attribute values change over time. However, in some domains, specific attribute value relationships are expected to remain consistent over a time-interval. For example, in therapeutic drug monitoring, the measurement of specific drug levels is expected to remain constant at defined time intervals. A patient must be given a drug dose within a time interval to ensure effective and therapeutic drug concentrations remain in the body. Existing work has largely focused on static graphs, and are unable to capture attribute value relationships over a time interval. Recent work has proposed a new class of dependencies, Temporal Graph Functional Dependencies (TGFDs) that model such semantics [Alipour et. al]. However, declarative specification of these constraints is challenging given the large-scale of real-life graphs. In this thesis, we propose, ParaTGFDMiner, a parallel discovery algorithm to identify TGFDs from large-scale graphs. Given an input graph G, ParaTGFDMiner discovers TGFDs by first generating candidate constraints locally at each worker node from a subset of G. A coordinator then validates and merges these candidate constraints, across all workers, to compute high-frequency TGFDs that are satisfied in G. We propose two optimizations that implement asynchronous processing to minimize idle time, and fast graph pattern matching to identify high-likelihood candidates. Our evaluation on two real large-scale graphs demonstrate that ParaTGFDMiner is up to 60% faster than baseline approaches.","abstract_html":"Temporal graphs play a crucial role in modelling dynamic applications where (node) attribute values change over time. However, in some domains, specific attribute value relationships are expected to remain consistent over a time-interval. For example, in therapeutic drug monitoring, the measurement of specific drug levels is expected to remain constant at defined time intervals. A patient must be given a drug dose within a time interval to ensure effective and therapeutic drug concentrations remain in the body. Existing work has largely focused on static graphs, and are unable to capture attribute value relationships over a time interval. Recent work has proposed a new class of dependencies, Temporal Graph Functional Dependencies (TGFDs) that model such semantics [Alipour et. al]. However, declarative specification of these constraints is challenging given the large-scale of real-life graphs. In this thesis, we propose, ParaTGFDMiner, a parallel discovery algorithm to identify TGFDs from large-scale graphs. Given an input graph G, ParaTGFDMiner discovers TGFDs by first generating candidate constraints locally at each worker node from a subset of G. A coordinator then validates and merges these candidate constraints, across all workers, to compute high-frequency TGFDs that are satisfied in G. We propose two optimizations that implement asynchronous processing to minimize idle time, and fast graph pattern matching to identify high-likelihood candidates. Our evaluation on two real large-scale graphs demonstrate that ParaTGFDMiner is up to 60% faster than baseline approaches.","abstract_has_math":false,"creators":["Wang, Yaoxu"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":"Computer Science","school":null,"contributors":[],"advisors":["Chiang, Fei"],"committee_chairs":[],"committee_members":[],"year":2024,"date_issued":"2024","date_published":"2024","updated_at":"2026-11-18T19:39:27Z","subjects":[],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[],"person_identifier_entries":[]},"links":{"outbound_url":"http://hdl.handle.net/11375/30264","outbound_label":"Handle","outbound_source":"dc:identifier.uri","record_url":"http://hdl.handle.net/11375/30264","record_label":"Repository record","record_kind":"handle","record_source":"dc:identifier.uri","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://macsphere.mcmaster.ca/server/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Amacsphere.mcmaster.ca%3A11375%2F30264","prefix":"dim"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Chiang, Fei"]},{"key":"dc:contributor.department","label":"Department","values":["Computer Science"]},{"key":"dc:creator","label":"Author","values":["Wang, Yaoxu"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2024-09-30T18:54:19Z"]},{"key":"dc:date.available","label":"Dc Date Available","values":["2024-09-30T18:54:19Z"]},{"key":"dc:date.issued","label":"Date","values":["2024"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["http://hdl.handle.net/11375/30264"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Temporal graphs play a crucial role in modelling dynamic applications where (node) attribute values change over time. However, in some domains, specific attribute value relationships are expected to remain consistent over a time-interval. For example, in therapeutic drug monitoring, the measurement of specific drug levels is expected to remain constant at defined time intervals. A patient must be given a drug dose within a time interval to ensure effective and therapeutic drug concentrations remain in the body. Existing work has largely focused on static graphs, and are unable to capture attribute value relationships over a time interval. Recent work has proposed a new class of dependencies, Temporal Graph Functional Dependencies (TGFDs) that model such semantics [Alipour et. al]. However, declarative specification of these constraints is challenging given the large-scale of real-life graphs. In this thesis, we propose, ParaTGFDMiner, a parallel discovery algorithm to identify TGFDs from large-scale graphs. Given an input graph G, ParaTGFDMiner discovers TGFDs by first generating candidate constraints locally at each worker node from a subset of G. A coordinator then validates and merges these candidate constraints, across all workers, to compute high-frequency TGFDs that are satisfied in G. We propose two optimizations that implement asynchronous processing to minimize idle time, and fast graph pattern matching to identify high-likelihood candidates. Our evaluation on two real large-scale graphs demonstrate that ParaTGFDMiner is up to 60% faster than baseline approaches."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Master of Science (MSc)"]},{"key":"dc:title","label":"Title","values":["Discovery of Temporal Graph Functional Dependencies"]},{"key":"others:access_status","label":"Others Access Status","values":["open.access"]}]}],"canonical_facts":{"dc:contributor.advisor":["Chiang, Fei"],"dc:contributor.department":["Computer Science"],"dc:creator":["Wang, Yaoxu"],"dc:date.accessioned":["2024-09-30T18:54:19Z"],"dc:date.available":["2024-09-30T18:54:19Z"],"dc:date.issued":["2024"],"dc:description.abstract":["Temporal graphs play a crucial role in modelling dynamic applications where (node) attribute values change over time. However, in some domains, specific attribute value relationships are expected to remain consistent over a time-interval. For example, in therapeutic drug monitoring, the measurement of specific drug levels is expected to remain constant at defined time intervals. A patient must be given a drug dose within a time interval to ensure effective and therapeutic drug concentrations remain in the body. Existing work has largely focused on static graphs, and are unable to capture attribute value relationships over a time interval. Recent work has proposed a new class of dependencies, Temporal Graph Functional Dependencies (TGFDs) that model such semantics [Alipour et. al]. However, declarative specification of these constraints is challenging given the large-scale of real-life graphs. In this thesis, we propose, ParaTGFDMiner, a parallel discovery algorithm to identify TGFDs from large-scale graphs. Given an input graph G, ParaTGFDMiner discovers TGFDs by first generating candidate constraints locally at each worker node from a subset of G. A coordinator then validates and merges these candidate constraints, across all workers, to compute high-frequency TGFDs that are satisfied in G. We propose two optimizations that implement asynchronous processing to minimize idle time, and fast graph pattern matching to identify high-likelihood candidates. Our evaluation on two real large-scale graphs demonstrate that ParaTGFDMiner is up to 60% faster than baseline approaches."],"dc:description.degree":["Master of Science (MSc)"],"dc:identifier.uri":["http://hdl.handle.net/11375/30264"],"dc:title":["Discovery of Temporal Graph Functional Dependencies"],"dc:type":["Thesis"],"others:access_status":["open.access"]},"updated_at":"2026-11-18T19:39:27Z"}