Effective mobile context pattern discovery via adapted hierarchical dirichlet processes

Jiangchuan Zheng, Siyuan Liu, Lionel M. Ni

Research output: Chapter in Book/Report/Conference proceedingConference contribution

7 Citations (Scopus)

Abstract

The extraction of macroscopic mobile context reflecting users' personal and social behavior patterns from smartphone sensor data (e.g., GPS/Bluetooth signals) is crucial in building intelligent pervasive systems. Hierarchical Dirichlet Processes (HDP), a well known Bayesian nonparametrics model for grouped data, is a promising option to achieve this objective due to its ability of discovering high-level semantics behind raw signals and establishing connections between individuals. However, applying HDP in a straightforward manner may not work as it does not take certain unique characteristics in mobile context into account. Particularly, while traditional HDP typically models a single aspect (e.g., Word), the characterization of a mobile context normally involves multiple heterogeneous aspects (e.g., Time, location, Bluetooth proximity). In addition, the presence of multiple aspects dictates a flexible way of clustering users and organizing mobile contexts in a hierarchical manner in serving different pervasive applications, a feature that traditional HDP lacks. Therefore, in this paper, we propose several extensions on traditional HDP to adapt it to the task of mobile context discovery. The key features in our extensions are: i) fusing multiple aspects naturally in HDP to achieve effective extraction of complex mobile context, ii) treating different aspects heterogeneously (globally or personally) in HDP to enable flexible user behavior clustering at various granularities in accordance with applications' needs, and iii) organizing mobile contexts in a hierarchical manner for natural behavior representation and overcoming data sparsity. Based on the experiments in a popular real-world mobile data set, we illustrate the ability of the framework in extracting useful mobile contexts such as characterizing personal life routines, discovering dominant temporal habits in a population, and inferring social group patterns, as well as its potential in improving individual mobility prediction under data sparsity.

Original languageEnglish (US)
Title of host publicationProceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages146-155
Number of pages10
Volume1
ISBN (Electronic)9781479957057
DOIs
StatePublished - Jan 1 2014
Event15th IEEE International Conference on Mobile Data Management, IEEE MDM 2014 - Brisbane, Australia
Duration: Jul 15 2014Jul 18 2014

Other

Other15th IEEE International Conference on Mobile Data Management, IEEE MDM 2014
CountryAustralia
CityBrisbane
Period7/15/147/18/14

Fingerprint

Bluetooth
Intelligent buildings
Smartphones
Global positioning system
Semantics
Sensors
Experiments

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Zheng, J., Liu, S., & Ni, L. M. (2014). Effective mobile context pattern discovery via adapted hierarchical dirichlet processes. In Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014 (Vol. 1, pp. 146-155). [6916915] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/MDM.2014.24
Zheng, Jiangchuan ; Liu, Siyuan ; Ni, Lionel M. / Effective mobile context pattern discovery via adapted hierarchical dirichlet processes. Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014. Vol. 1 Institute of Electrical and Electronics Engineers Inc., 2014. pp. 146-155
@inproceedings{242d34c83a644f359d342a255f11f89b,
title = "Effective mobile context pattern discovery via adapted hierarchical dirichlet processes",
abstract = "The extraction of macroscopic mobile context reflecting users' personal and social behavior patterns from smartphone sensor data (e.g., GPS/Bluetooth signals) is crucial in building intelligent pervasive systems. Hierarchical Dirichlet Processes (HDP), a well known Bayesian nonparametrics model for grouped data, is a promising option to achieve this objective due to its ability of discovering high-level semantics behind raw signals and establishing connections between individuals. However, applying HDP in a straightforward manner may not work as it does not take certain unique characteristics in mobile context into account. Particularly, while traditional HDP typically models a single aspect (e.g., Word), the characterization of a mobile context normally involves multiple heterogeneous aspects (e.g., Time, location, Bluetooth proximity). In addition, the presence of multiple aspects dictates a flexible way of clustering users and organizing mobile contexts in a hierarchical manner in serving different pervasive applications, a feature that traditional HDP lacks. Therefore, in this paper, we propose several extensions on traditional HDP to adapt it to the task of mobile context discovery. The key features in our extensions are: i) fusing multiple aspects naturally in HDP to achieve effective extraction of complex mobile context, ii) treating different aspects heterogeneously (globally or personally) in HDP to enable flexible user behavior clustering at various granularities in accordance with applications' needs, and iii) organizing mobile contexts in a hierarchical manner for natural behavior representation and overcoming data sparsity. Based on the experiments in a popular real-world mobile data set, we illustrate the ability of the framework in extracting useful mobile contexts such as characterizing personal life routines, discovering dominant temporal habits in a population, and inferring social group patterns, as well as its potential in improving individual mobility prediction under data sparsity.",
author = "Jiangchuan Zheng and Siyuan Liu and Ni, {Lionel M.}",
year = "2014",
month = "1",
day = "1",
doi = "10.1109/MDM.2014.24",
language = "English (US)",
volume = "1",
pages = "146--155",
booktitle = "Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

Zheng, J, Liu, S & Ni, LM 2014, Effective mobile context pattern discovery via adapted hierarchical dirichlet processes. in Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014. vol. 1, 6916915, Institute of Electrical and Electronics Engineers Inc., pp. 146-155, 15th IEEE International Conference on Mobile Data Management, IEEE MDM 2014, Brisbane, Australia, 7/15/14. https://doi.org/10.1109/MDM.2014.24

Effective mobile context pattern discovery via adapted hierarchical dirichlet processes. / Zheng, Jiangchuan; Liu, Siyuan; Ni, Lionel M.

Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014. Vol. 1 Institute of Electrical and Electronics Engineers Inc., 2014. p. 146-155 6916915.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Effective mobile context pattern discovery via adapted hierarchical dirichlet processes

AU - Zheng, Jiangchuan

AU - Liu, Siyuan

AU - Ni, Lionel M.

PY - 2014/1/1

Y1 - 2014/1/1

N2 - The extraction of macroscopic mobile context reflecting users' personal and social behavior patterns from smartphone sensor data (e.g., GPS/Bluetooth signals) is crucial in building intelligent pervasive systems. Hierarchical Dirichlet Processes (HDP), a well known Bayesian nonparametrics model for grouped data, is a promising option to achieve this objective due to its ability of discovering high-level semantics behind raw signals and establishing connections between individuals. However, applying HDP in a straightforward manner may not work as it does not take certain unique characteristics in mobile context into account. Particularly, while traditional HDP typically models a single aspect (e.g., Word), the characterization of a mobile context normally involves multiple heterogeneous aspects (e.g., Time, location, Bluetooth proximity). In addition, the presence of multiple aspects dictates a flexible way of clustering users and organizing mobile contexts in a hierarchical manner in serving different pervasive applications, a feature that traditional HDP lacks. Therefore, in this paper, we propose several extensions on traditional HDP to adapt it to the task of mobile context discovery. The key features in our extensions are: i) fusing multiple aspects naturally in HDP to achieve effective extraction of complex mobile context, ii) treating different aspects heterogeneously (globally or personally) in HDP to enable flexible user behavior clustering at various granularities in accordance with applications' needs, and iii) organizing mobile contexts in a hierarchical manner for natural behavior representation and overcoming data sparsity. Based on the experiments in a popular real-world mobile data set, we illustrate the ability of the framework in extracting useful mobile contexts such as characterizing personal life routines, discovering dominant temporal habits in a population, and inferring social group patterns, as well as its potential in improving individual mobility prediction under data sparsity.

AB - The extraction of macroscopic mobile context reflecting users' personal and social behavior patterns from smartphone sensor data (e.g., GPS/Bluetooth signals) is crucial in building intelligent pervasive systems. Hierarchical Dirichlet Processes (HDP), a well known Bayesian nonparametrics model for grouped data, is a promising option to achieve this objective due to its ability of discovering high-level semantics behind raw signals and establishing connections between individuals. However, applying HDP in a straightforward manner may not work as it does not take certain unique characteristics in mobile context into account. Particularly, while traditional HDP typically models a single aspect (e.g., Word), the characterization of a mobile context normally involves multiple heterogeneous aspects (e.g., Time, location, Bluetooth proximity). In addition, the presence of multiple aspects dictates a flexible way of clustering users and organizing mobile contexts in a hierarchical manner in serving different pervasive applications, a feature that traditional HDP lacks. Therefore, in this paper, we propose several extensions on traditional HDP to adapt it to the task of mobile context discovery. The key features in our extensions are: i) fusing multiple aspects naturally in HDP to achieve effective extraction of complex mobile context, ii) treating different aspects heterogeneously (globally or personally) in HDP to enable flexible user behavior clustering at various granularities in accordance with applications' needs, and iii) organizing mobile contexts in a hierarchical manner for natural behavior representation and overcoming data sparsity. Based on the experiments in a popular real-world mobile data set, we illustrate the ability of the framework in extracting useful mobile contexts such as characterizing personal life routines, discovering dominant temporal habits in a population, and inferring social group patterns, as well as its potential in improving individual mobility prediction under data sparsity.

UR - http://www.scopus.com/inward/record.url?scp=84907986956&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907986956&partnerID=8YFLogxK

U2 - 10.1109/MDM.2014.24

DO - 10.1109/MDM.2014.24

M3 - Conference contribution

AN - SCOPUS:84907986956

VL - 1

SP - 146

EP - 155

BT - Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014

PB - Institute of Electrical and Electronics Engineers Inc.

ER -

Zheng J, Liu S, Ni LM. Effective mobile context pattern discovery via adapted hierarchical dirichlet processes. In Proceedings - 2014 IEEE 15th International Conference on Mobile Data Management, IEEE MDM 2014. Vol. 1. Institute of Electrical and Electronics Engineers Inc. 2014. p. 146-155. 6916915 https://doi.org/10.1109/MDM.2014.24