{"id":{"repo_id":"uiuc","oai_identifier":"oai:www.ideals.illinois.edu:2142/87405"},"canonical_url":"https://search.dev.ndltd.org/etd/uiuc/oai:www.ideals.illinois.edu:2142/87405","repository":{"repo_id":"uiuc","name":"University of Illinois - Urbana-Champaign","base_url":"https://www.ideals.illinois.edu/oai-pmh"},"display":{"title":"Clustering Analysis for Non-Stationary Time Series","abstract":"77 p.","abstract_html":"77 p.","abstract_has_math":false,"creators":["Gao, Bing"],"institution":"University of Illinois at Urbana-Champaign","degree_name":"Ph.D.","degree_level":"Dissertation","degree_discipline":"Statistics","degree_department":null,"school":null,"contributors":["Hernando Ombao"],"advisors":[],"committee_chairs":[],"committee_members":[],"year":2015,"date_issued":"2015-09-28T16:02:44Z","date_published":"2015-09-28T16:02:44Z","updated_at":"2026-07-22T22:26:30Z","subjects":["Statistics"],"languages":["eng"],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier","label":"Identifier","values":["(MiAaPQ)AAI3242843"],"render_values":[{"text":"(MiAaPQ)AAI3242843","href":null,"code":true}]}]},"links":{"outbound_url":"http://hdl.handle.net/2142/87405","outbound_label":"Handle","outbound_source":"dc:identifier"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor","label":"Contributor","values":["Hernando Ombao"]},{"key":"dc:creator","label":"Author","values":["Gao, Bing"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date","label":"Dc Date","values":["2015-09-28T16:02:44Z","10000-01-01","2006"]},{"key":"dc:type","label":"Dc Type","values":["text"]},{"key":"thesis:degree_discipline","label":"Discipline","values":["Statistics"]},{"key":"thesis:degree_level","label":"Degree Level","values":["Dissertation"]},{"key":"thesis:degree_name","label":"Degree Name","values":["Ph.D."]},{"key":"thesis:institution_name","label":"Thesis Institution Name","values":["University of Illinois at Urbana-Champaign"]}]},{"id":"subjects_keywords","label":"Subjects and Keywords","entries":[{"key":"dc:subject","label":"Dc Subject","values":["Statistics"]}]},{"id":"language_rights","label":"Language and Rights","entries":[{"key":"dc:language","label":"Dc Language","values":["eng"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier","label":"Identifier","values":["http://hdl.handle.net/2142/87405","(MiAaPQ)AAI3242843"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description","label":"Description","values":["77 p.","The classical approaches to clustering are hierarchical and k-means. They are popular in practice. However, they can not address the issue of determining the number of clusters within the data. In this dissertation, we develop a best model-based clustering algorithm that can automatically select the best features for clustering, and estimate the number of clusters. The whole procedure can be divided into 2 steps. The first step is to find a basis from the WPs library that can best illuminate the difference among groups of the time series. The basis selected will consist of T WP functions, many of them irrelevant for discriminating/clustering groups. Thus, in the second step, we use the model-based variable selection algorithm to select the WPs that are really useful for clustering from the basis chosen in the first step. Redo the above 2 steps for each G from 2 to a preselect maximum number of clusters. Finally compare the models selected for each G by BIC. The best model contains information about the number of clusters and cluster membership. Moreover, based on best G, and the best WPs selected from the best basis, the EM result provides a measure of uncertainty about the associated classification of each time series. Simulation studies have been carried out and demonstrated that our method works well. We have also applied the method to a seismic data set, cluster the signals as earthquakes or explosions, and to epileptic seizure EEG data, separate the signals before and during epileptic seizure.","Made available in DSpace on 2015-09-28T16:02:44Z (GMT). No. of bitstreams: 2 license.txt: 4848 bytes, checksum: 96035ab3f5e1c23cc7138a224ce498bd (MD5) 3242843.pdf: 1885753 bytes, checksum: dfbcd2998790a6f13be652c454bf215d (MD5) Previous issue date: 2006","Embargo set by: Seth Robbins for item 88686 Lift date: Forever Reason: Restricted to the U of I community idenfinitely during batch ingest of legacy ETDs","Restricted to the U of I community idenfinitely during batch ingest of legacy ETDs","U of I Only","Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2006."]},{"key":"dc:title","label":"Title","values":["Clustering Analysis for Non-Stationary Time Series"]}]}],"canonical_facts":{"dc:contributor":["Hernando Ombao"],"dc:creator":["Gao, Bing"],"dc:date":["2015-09-28T16:02:44Z","10000-01-01","2006"],"dc:description":["77 p.","The classical approaches to clustering are hierarchical and k-means. They are popular in practice. However, they can not address the issue of determining the number of clusters within the data. In this dissertation, we develop a best model-based clustering algorithm that can automatically select the best features for clustering, and estimate the number of clusters. The whole procedure can be divided into 2 steps. The first step is to find a basis from the WPs library that can best illuminate the difference among groups of the time series. The basis selected will consist of T WP functions, many of them irrelevant for discriminating/clustering groups. Thus, in the second step, we use the model-based variable selection algorithm to select the WPs that are really useful for clustering from the basis chosen in the first step. Redo the above 2 steps for each G from 2 to a preselect maximum number of clusters. Finally compare the models selected for each G by BIC. The best model contains information about the number of clusters and cluster membership. Moreover, based on best G, and the best WPs selected from the best basis, the EM result provides a measure of uncertainty about the associated classification of each time series. Simulation studies have been carried out and demonstrated that our method works well. We have also applied the method to a seismic data set, cluster the signals as earthquakes or explosions, and to epileptic seizure EEG data, separate the signals before and during epileptic seizure.","Made available in DSpace on 2015-09-28T16:02:44Z (GMT). No. of bitstreams: 2 license.txt: 4848 bytes, checksum: 96035ab3f5e1c23cc7138a224ce498bd (MD5) 3242843.pdf: 1885753 bytes, checksum: dfbcd2998790a6f13be652c454bf215d (MD5) Previous issue date: 2006","Embargo set by: Seth Robbins for item 88686 Lift date: Forever Reason: Restricted to the U of I community idenfinitely during batch ingest of legacy ETDs","Restricted to the U of I community idenfinitely during batch ingest of legacy ETDs","U of I Only","Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 2006."],"dc:identifier":["http://hdl.handle.net/2142/87405","(MiAaPQ)AAI3242843"],"dc:language":["eng"],"dc:subject":["Statistics"],"dc:title":["Clustering Analysis for Non-Stationary Time Series"],"dc:type":["text"],"thesis:degree_discipline":["Statistics"],"thesis:degree_level":["Dissertation"],"thesis:degree_name":["Ph.D."],"thesis:institution_name":["University of Illinois at Urbana-Champaign"]},"updated_at":"2026-07-22T22:26:30Z"}