FRINK Linked Data Fragments server

Matches in SemOpenAlex for { <https://semopenalex.org/work/W2547083798> ?p ?o ?g. }

• W2547083798 endingPage "225" @default.
• W2547083798 startingPage "225" @default.
• W2547083798 abstract "This dissertation focuses on pattern discovery in spatial, image, and biological data and addresses the problem from the perspective of data reduction. We consider three kinds of reduction that can directly or indirectly help pattern discovery: noise removal, compression, and clustering. A graph-based noise removing method is proposed and demonstrated superior over traditional distance-based methods on pattern quality. Two visualization models are devised to assist and evaluate different noise removing methods on parameter selection and encode user requirements. In addition, a compression method is proposed to condense a given dataset into a compact version on which mining operations can be performed efficiently. The compression process can terminate automatically and the result is deterministic and reproducible. Finally, the proposed noise removal and compression can be synthesized with a hierarchical merging criterion to discover clusters of different shapes, sizes, and densities in noisy spatial data efficiently. The resulting system can be found at http://viscomp.utdallas.edu/FACADE.The proposed reduction methods can be adapted to post-classification of remote sensing imagery. Experimental comparisons show that the adapted approach preserves more details of region boundary than existing methods. For certain types of images with similar content, a learning-based approach is proposed for parameter selection so that automatic image recognition becomes possible. Besides, we investigate the possibility of encoding more user constraints into the process of cluster analysis within a graph model and a set of criteria and corresponding solving algorithms are presented.The final part of this dissertation reports preliminary results of high-dimensional FACS data analysis and proposes a grid-based clustering approach based on the analysis to identify the invisible cell populations from huge amount of flow cytometry data. The proposed clustering approach can not only identify number of clusters automatically but also be scalable on both data dimension and size. Preliminary experimental results show that our approach reduces computation time over the traditional k-means clustering method by more than an order of magnitude and produces better results without needing prior knowledge from users. Compared with FlowJo, the popular commercial software of flow cytometry analysis, our approach is also superior on speed and number of user inputs." @default.
• W2547083798 created "2016-11-11" @default.
• W2547083798 creator A5029636169 @default.
• W2547083798 creator A5034262562 @default.
• W2547083798 creator A5071258270 @default.
• W2547083798 date "2006-01-01" @default.
• W2547083798 modified "2023-09-25" @default.
• W2547083798 title "Pattern discovery in spatial, image, and biological data" @default.
• W2547083798 hasPublicationYear "2006" @default.
• W2547083798 type Work @default.
• W2547083798 sameAs 2547083798 @default.
• W2547083798 citedByCount "0" @default.
• W2547083798 crossrefType "journal-article" @default.
• W2547083798 hasAuthorship W2547083798A5029636169 @default.
• W2547083798 hasAuthorship W2547083798A5034262562 @default.
• W2547083798 hasAuthorship W2547083798A5071258270 @default.
• W2547083798 hasConcept C115961682 @default.
• W2547083798 hasConcept C124101348 @default.
• W2547083798 hasConcept C132525143 @default.
• W2547083798 hasConcept C13481523 @default.
• W2547083798 hasConcept C153180895 @default.
• W2547083798 hasConcept C154945302 @default.
• W2547083798 hasConcept C163294075 @default.
• W2547083798 hasConcept C36464697 @default.
• W2547083798 hasConcept C41008148 @default.
• W2547083798 hasConcept C73555534 @default.
• W2547083798 hasConcept C78548338 @default.
• W2547083798 hasConcept C80444323 @default.
• W2547083798 hasConcept C9417928 @default.
• W2547083798 hasConcept C99498987 @default.
• W2547083798 hasConceptScore W2547083798C115961682 @default.
• W2547083798 hasConceptScore W2547083798C124101348 @default.
• W2547083798 hasConceptScore W2547083798C132525143 @default.
• W2547083798 hasConceptScore W2547083798C13481523 @default.
• W2547083798 hasConceptScore W2547083798C153180895 @default.
• W2547083798 hasConceptScore W2547083798C154945302 @default.
• W2547083798 hasConceptScore W2547083798C163294075 @default.
• W2547083798 hasConceptScore W2547083798C36464697 @default.
• W2547083798 hasConceptScore W2547083798C41008148 @default.
• W2547083798 hasConceptScore W2547083798C73555534 @default.
• W2547083798 hasConceptScore W2547083798C78548338 @default.
• W2547083798 hasConceptScore W2547083798C80444323 @default.
• W2547083798 hasConceptScore W2547083798C9417928 @default.
• W2547083798 hasConceptScore W2547083798C99498987 @default.
• W2547083798 hasLocation W25470837981 @default.
• W2547083798 hasOpenAccess W2547083798 @default.
• W2547083798 hasPrimaryLocation W25470837981 @default.
• W2547083798 hasRelatedWork W2030993905 @default.
• W2547083798 hasRelatedWork W2053628764 @default.
• W2547083798 hasRelatedWork W2090286885 @default.
• W2547083798 hasRelatedWork W2315191261 @default.
• W2547083798 hasRelatedWork W2330745922 @default.
• W2547083798 hasRelatedWork W2394718665 @default.
• W2547083798 hasRelatedWork W2402944866 @default.
• W2547083798 hasRelatedWork W2411976423 @default.
• W2547083798 hasRelatedWork W2500440664 @default.
• W2547083798 hasRelatedWork W2573076041 @default.
• W2547083798 hasRelatedWork W2800881133 @default.
• W2547083798 hasRelatedWork W2802597471 @default.
• W2547083798 hasRelatedWork W2889794534 @default.
• W2547083798 hasRelatedWork W2902533660 @default.
• W2547083798 hasRelatedWork W2902853028 @default.
• W2547083798 hasRelatedWork W2963901276 @default.
• W2547083798 hasRelatedWork W3015626561 @default.
• W2547083798 hasRelatedWork W3023858501 @default.
• W2547083798 hasRelatedWork W3149863265 @default.
• W2547083798 hasRelatedWork W3207109271 @default.
• W2547083798 isParatext "false" @default.
• W2547083798 isRetracted "false" @default.
• W2547083798 magId "2547083798" @default.
• W2547083798 workType "article" @default.