{"id":{"repo_id":"msu","oai_identifier":"oai:d.lib.msu.edu:etd_50534"},"canonical_url":"https://search.ndltd.org/etd/msu/oai:d.lib.msu.edu:etd_50534","repository":{"repo_id":"msu","name":"Michigan State University","base_url":"https://d.lib.msu.edu/oai"},"display":{"title":"Using the deep graph representation learning method to study the Indian election social network","abstract":"The aim of this dissertation is using the Deep Graph Learning to study on two problems in social network data. This dissertation includes two essays. The first essay is using Deep Graph Representation Learning to solve the counterfactual inference problem on social network. The second essay is using Deep Graph Representation Learning to solve the fairness recommendation results on social network.Graph Representation Learning is very useful method to extract graph features from the large-scale network. There are many advantages when we use this method to map the data from the graph space into the embedding space. First, there is no need to use the predefined measurement on graph structure such as bipartite graph, tripartite graph, etc, but directly keep the information of the network topology as good as possible. Second, when we study on social network platform such as Facebook, Twitter, etc, we need to consider more attributes with nodes and edges. Therefore, the information in the social network becomes more complicated. Considering these two advantages, we choose the graph representation learning to embed these features into the representation space.The first essay is combining this method with the transfer learning in order to fit the Rubin's counterfactual framework. It can help to give a robust and lowest biased estimation results compared with the propensity score matching methods.The second essay is borrowing the advantage of the graph representation learning method to learn the representation space of different types of the users and connections in twitter. We combine this method with the attention mechanism to construct the fairness based loss function. This can help to increase the fairness of the recommendation and maintain the predicted accuracy of the recommendation in social network.","abstract_html":"The aim of this dissertation is using the Deep Graph Learning to study on two problems in social network data. This dissertation includes two essays. The first essay is using Deep Graph Representation Learning to solve the counterfactual inference problem on social network. The second essay is using Deep Graph Representation Learning to solve the fairness recommendation results on social network.Graph Representation Learning is very useful method to extract graph features from the large-scale network. There are many advantages when we use this method to map the data from the graph space into the embedding space. First, there is no need to use the predefined measurement on graph structure such as bipartite graph, tripartite graph, etc, but directly keep the information of the network topology as good as possible. Second, when we study on social network platform such as Facebook, Twitter, etc, we need to consider more attributes with nodes and edges. Therefore, the information in the social network becomes more complicated. Considering these two advantages, we choose the graph representation learning to embed these features into the representation space.The first essay is combining this method with the transfer learning in order to fit the Rubin&#x27;s counterfactual framework. It can help to give a robust and lowest biased estimation results compared with the propensity score matching methods.The second essay is borrowing the advantage of the graph representation learning method to learn the representation space of different types of the users and connections in twitter. We combine this method with the attention mechanism to construct the fairness based loss function. This can help to increase the fairness of the recommendation and maintain the predicted accuracy of the recommendation in social network.","abstract_has_math":false,"creators":["Liu, Yimo"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":null,"school":null,"contributors":["Susarla, Anjana","Zhang, Quan","Xie, Yuying","Hirn, Matthew"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2022,"date_issued":"2022","date_published":"2022","updated_at":"2026-11-18T18:57:00Z","subjects":["Industrial management","Graph theory--Data processing","Counterfactuals (Logic)","Data mining","Social networks"],"languages":["English"],"rights":["In Copyright"],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:50534","isbn:9798834045762","oclc:on1373601434","oclc:1373601434","umi:29255686","local:Liu_grad.msu_0128D_19169"],"render_values":[{"text":"etd:50534","href":null,"code":true},{"text":"isbn:9798834045762","href":null,"code":true},{"text":"oclc:on1373601434","href":null,"code":true},{"text":"oclc:1373601434","href":null,"code":true},{"text":"umi:29255686","href":null,"code":true},{"text":"local:Liu_grad.msu_0128D_19169","href":null,"code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://doi.org/doi:10.25335/yfd0-zw75","outbound_label":"DOI","outbound_source":"dc:identifier","record_url":"https://doi.org/doi:10.25335/yfd0-zw75","record_label":"Repository record","record_kind":"doi","record_source":"dc:identifier","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://d.lib.msu.edu/oai?verb=GetRecord&metadataPrefix=oai_dc&identifier=oai%3Ad.lib.msu.edu%3Aetd_50534","prefix":"oai_dc"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Susarla, Anjana","Zhang, Quan","Xie, Yuying","Hirn, Matthew"]},{"key":"dc:creator","label":"Author","values":["Liu, Yimo"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2022"]},{"key":"dc:relation","label":"Dc Relation","values":["Electronic Theses & Dissertations"]},{"key":"dc:type","label":"Dc Type","values":["Text","Theses"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Industrial management","Graph theory--Data processing","Counterfactuals (Logic)","Data mining","Social networks"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["English"]},{"key":"dc:rights","label":"Dc Rights","values":["In Copyright"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["etd:50534","isbn:9798834045762","oclc:on1373601434","oclc:1373601434","umi:29255686","local:Liu_grad.msu_0128D_19169","https://doi.org/doi:10.25335/yfd0-zw75"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["The aim of this dissertation is using the Deep Graph Learning to study on two problems in social network data. This dissertation includes two essays. The first essay is using Deep Graph Representation Learning to solve the counterfactual inference problem on social network. The second essay is using Deep Graph Representation Learning to solve the fairness recommendation results on social network.Graph Representation Learning is very useful method to extract graph features from the large-scale network. There are many advantages when we use this method to map the data from the graph space into the embedding space. First, there is no need to use the predefined measurement on graph structure such as bipartite graph, tripartite graph, etc, but directly keep the information of the network topology as good as possible. Second, when we study on social network platform such as Facebook, Twitter, etc, we need to consider more attributes with nodes and edges. Therefore, the information in the social network becomes more complicated. Considering these two advantages, we choose the graph representation learning to embed these features into the representation space.The first essay is combining this method with the transfer learning in order to fit the Rubin's counterfactual framework. It can help to give a robust and lowest biased estimation results compared with the propensity score matching methods.The second essay is borrowing the advantage of the graph representation learning method to learn the representation space of different types of the users and connections in twitter. We combine this method with the attention mechanism to construct the fairness based loss function. This can help to increase the fairness of the recommendation and maintain the predicted accuracy of the recommendation in social network.","Thesis (Ph. D.)--Michigan State University. Business Administration, 2022","Includes bibliographical references (pages 151-163)"]},{"key":"dc:format","label":"Dc Format","values":["xii, 163 pages","application/pdf"]},{"key":"dc:title","label":"Title","values":["Using the deep graph representation learning method to study the Indian election social network"]}]}],"canonical_facts":{"dc:contributor":["Susarla, Anjana","Zhang, Quan","Xie, Yuying","Hirn, Matthew"],"dc:creator":["Liu, Yimo"],"dc:date":["2022"],"dc:description":["The aim of this dissertation is using the Deep Graph Learning to study on two problems in social network data. This dissertation includes two essays. The first essay is using Deep Graph Representation Learning to solve the counterfactual inference problem on social network. The second essay is using Deep Graph Representation Learning to solve the fairness recommendation results on social network.Graph Representation Learning is very useful method to extract graph features from the large-scale network. There are many advantages when we use this method to map the data from the graph space into the embedding space. First, there is no need to use the predefined measurement on graph structure such as bipartite graph, tripartite graph, etc, but directly keep the information of the network topology as good as possible. Second, when we study on social network platform such as Facebook, Twitter, etc, we need to consider more attributes with nodes and edges. Therefore, the information in the social network becomes more complicated. Considering these two advantages, we choose the graph representation learning to embed these features into the representation space.The first essay is combining this method with the transfer learning in order to fit the Rubin's counterfactual framework. It can help to give a robust and lowest biased estimation results compared with the propensity score matching methods.The second essay is borrowing the advantage of the graph representation learning method to learn the representation space of different types of the users and connections in twitter. We combine this method with the attention mechanism to construct the fairness based loss function. This can help to increase the fairness of the recommendation and maintain the predicted accuracy of the recommendation in social network.","Thesis (Ph. D.)--Michigan State University. Business Administration, 2022","Includes bibliographical references (pages 151-163)"],"dc:format":["xii, 163 pages","application/pdf"],"dc:identifier":["etd:50534","isbn:9798834045762","oclc:on1373601434","oclc:1373601434","umi:29255686","local:Liu_grad.msu_0128D_19169","https://doi.org/doi:10.25335/yfd0-zw75"],"dc:language":["English"],"dc:relation":["Electronic Theses & Dissertations"],"dc:rights":["In Copyright"],"dc:subject":["Industrial management","Graph theory--Data processing","Counterfactuals (Logic)","Data mining","Social networks"],"dc:title":["Using the deep graph representation learning method to study the Indian election social network"],"dc:type":["Text","Theses"]},"updated_at":"2026-11-18T18:57:00Z"}