東華大學圖書館 |

Language: English

Help

回圖書館首頁

手機版館藏查詢

Back

Switch To: Labeled | MARC Mode | ISBD

Linked to FindBook

Google Book

Amazon

博客來

Towards Multimodal Spatiotemporal Data Analysis : = Heterogeneity and Fusion.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Towards Multimodal Spatiotemporal Data Analysis :/
Reminder of title:	Heterogeneity and Fusion.
Author:	Zhang, Yawen.
Description:	1 online resource (147 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 84-03, Section: B.
Contained By:	Dissertations Abstracts International84-03B.
Subject:	Computer science. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29253522click for full text (PQDT)
ISBN:	9798841798385

Towards Multimodal Spatiotemporal Data Analysis : = Heterogeneity and Fusion.
Zhang, Yawen.

Towards Multimodal Spatiotemporal Data Analysis :Heterogeneity and Fusion. - 1 online resource (147 pages)

Source: Dissertations Abstracts International, Volume: 84-03, Section: B.

Thesis (Ph.D.)--University of Colorado at Boulder, 2022.

Includes bibliographical references

Spatiaotemporal data are data with space and/or time dimensions. With the widespread use of Internet-of-Things (IoT) technology and deployment of different types of sensors, either mobile or static, large volumes of spatiotemporal data are continuously collected and processed, and have become ubiquitous in the real world. Various services and insights can be delivered via spatiotemporal data, ranging from earthquake early warning, air pollution monitoring and forecasting, to automatic identification of fish schools. A key challenge for deriving information and knowledge from spatiotemporal data lies in handling data heterogeneity. There are various types of heterogeneity associated with spatiotemporal data. The same type of data streams collected may demonstrate different qualities, and multi-source spatiotemporal data often demonstrate different structures, dimensionality, and spatial and temporal resolutions. For spatiotemporal applications, understanding and handling the problem of data heterogeneity are of significant importance. When leveraging multi-source or multimodal data for specific application, a key challenge lies in designing appropriate data fusion techniques to effectively integrate information from spatiotemporal data. To address the above challenges, this dissertation investigates the heterogeneity problem in spatiotemporal data, and proposes fusion methods to integrate information from multimodal spatiotemporal data. We conduct experiments in three applications: (1) analyzing sensing heterogeneity in a global smartphone-based seismic network, (2) fusing multi-source heterogeneous data for air pollution hotspot identification and pollution level prediction, and (3) integrating spatial and temporal information for fish school identification. In application (1), we systematically analyze accelerometer-based sensing quality using millions of acceleration data from the MyShake devices, and investigate various factors that may impact waveform data quality. In application (2), we propose a two-step approach to detect hotspots from mobile sensing data, and leverage cross-domain urban data for hotspot inference. We also propose multi-group Encoder-Decoder networks (MGED-Net) to effectively fuse multi-source data for next-day air quality prediction, which outperforms multiple baseline models. In application (3), we leverage a multi-view learning technique called co-training to integrate contextual information into the classification model, which effectively improve the accuracy of Atlantic herring school identification. We also propose a superpixel-based spatio-temporal contrastive learning method (SSTC) to generate effective representation for fish schools in an unsupervised manner, which outperforms multiple baseline models in two different classification tasks.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9798841798385Subjects--Topical Terms:

523869
Computer science.
Subjects--Index Terms:

Air qualityIndex Terms--Genre/Form:

542853
Electronic books.

Towards Multimodal Spatiotemporal Data Analysis : = Heterogeneity and Fusion.
LDR:04184nmm a2200385K 4500 001 2354877
005 20230505090447.5
006 m o d
007 cr mn ---uuuuu
008 241011s2022 xx obm 000 0 eng d
020 $a 9798841798385
035 $a (MiAaPQ)AAI29253522
035 $a AAI29253522
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Zhang, Yawen. $3 3681746
245 1 0 $a Towards Multimodal Spatiotemporal Data Analysis : $b Heterogeneity and Fusion.
264 0 $c 2022
300 $a 1 online resource (147 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 84-03, Section: B.
500 $a Advisor: Lv, Qin.
502 $a Thesis (Ph.D.)--University of Colorado at Boulder, 2022.
504 $a Includes bibliographical references
520 $a Spatiaotemporal data are data with space and/or time dimensions. With the widespread use of Internet-of-Things (IoT) technology and deployment of different types of sensors, either mobile or static, large volumes of spatiotemporal data are continuously collected and processed, and have become ubiquitous in the real world. Various services and insights can be delivered via spatiotemporal data, ranging from earthquake early warning, air pollution monitoring and forecasting, to automatic identification of fish schools. A key challenge for deriving information and knowledge from spatiotemporal data lies in handling data heterogeneity. There are various types of heterogeneity associated with spatiotemporal data. The same type of data streams collected may demonstrate different qualities, and multi-source spatiotemporal data often demonstrate different structures, dimensionality, and spatial and temporal resolutions. For spatiotemporal applications, understanding and handling the problem of data heterogeneity are of significant importance. When leveraging multi-source or multimodal data for specific application, a key challenge lies in designing appropriate data fusion techniques to effectively integrate information from spatiotemporal data. To address the above challenges, this dissertation investigates the heterogeneity problem in spatiotemporal data, and proposes fusion methods to integrate information from multimodal spatiotemporal data. We conduct experiments in three applications: (1) analyzing sensing heterogeneity in a global smartphone-based seismic network, (2) fusing multi-source heterogeneous data for air pollution hotspot identification and pollution level prediction, and (3) integrating spatial and temporal information for fish school identification. In application (1), we systematically analyze accelerometer-based sensing quality using millions of acceleration data from the MyShake devices, and investigate various factors that may impact waveform data quality. In application (2), we propose a two-step approach to detect hotspots from mobile sensing data, and leverage cross-domain urban data for hotspot inference. We also propose multi-group Encoder-Decoder networks (MGED-Net) to effectively fuse multi-source data for next-day air quality prediction, which outperforms multiple baseline models. In application (3), we leverage a multi-view learning technique called co-training to integrate contextual information into the classification model, which effectively improve the accuracy of Atlantic herring school identification. We also propose a superpixel-based spatio-temporal contrastive learning method (SSTC) to generate effective representation for fish schools in an unsupervised manner, which outperforms multiple baseline models in two different classification tasks.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Computer science. $3 523869
653 $a Air quality
653 $a Data fusion
653 $a Data heterogeneity
653 $a Earthquake
653 $a Fisheries acoustics
653 $a Spatiotemporal data
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0984
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a University of Colorado at Boulder. $b Computer Science. $3 1018560
773 0 $t Dissertations Abstracts International $g 84-03B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=29253522 $z click for full text (PQDT)