MobiGroup: Enabling Lifecycle Support to Social Activity Organization and Suggestion With Mobile Crowd Sensing

[pdf-embedder url=”https://wellapets.com/wp-content/uploads/2019/06/MobiGroup-Enabling-Lifecycle-Support-to-Social.pdf” title=”MobiGroup Enabling Lifecycle Support to Social”]

Abstract—This paper presents a group-aware mobile crowd
sensing system called MobiGroup, which supports group activity
organization in real-world settings. Acknowledging the complexity
and diversity of group activities, this paper introduces a formal
concept model to characterize group activities and classifies
them into four organizational stages. We then present an intelligent
approach to support group activity preparation, including a
heuristic rule-based mechanism for advertising public activity and
a context-based method for private group formation. In addition,
we leverage features extracted from both online and offline communities
to recommend ongoing events to attendees with different
needs. Compared with the baseline method, people preferred public
activities suggested by our heuristic rule-based method. Using a
dataset collected from 45 participants, we found that the contextbased
approach for private group formation can attain a precision
and recall of over 80%, and the usage of spatial–temporal contexts
and group computing can have more than a 30% performance improvement
over considering the interaction frequency between a
user and related groups. A case study revealed that, by extracting
the features such as dynamic intimacy and static intimacy, our
cross-community approach for ongoing event recommendation can
meet different user needs.
Index Terms—Cross-community sensing and mining, group
computing, mobile crowd sensing (MCS), social activity
organization.

INTRODUCTION
MANY technologies facilitate group interaction [1]. For
example, group management tools [2]–[6] can help analyze
historical interaction data of online communication (e.g.,
email and Facebook) and offline colocated social events [7]–
[11]. However, group activities involve more complex processes such as group formation and event publicity in addition to intragroup
interaction during the events.
This is a challenging problem. First of all, different types of
group activities may vary in goals, needs, constraints, flows, organizations,
and interaction patterns. For instance, some events
are closed or private (e.g., a party). In other cases, activities
are open to the public. In addition, each organizational stage of
an activity may require different technical support. For example,
one challenge of activity preparation is to locate and invite
potential attendees. In contrast, the core issue of running an
activity is recognizing and monitoring ongoing events. Group
would benefit from a conceptual model for automatic processing.
Second, there is a lack of technical infrastructure for group
activity logging and mining. For online communities, social web
portals can capture virtual interaction data for further use, such
as social tie detection. Data from real-world group activities are
harder to obtain, requiring specialized models, methods, and
mechanisms. It is more difficult to extract information and infer
knowledge from physical group activities, since the data tend
to be noisy and incomplete. Third, activity organization in realworld
settings is often influenced by various social and physical
contextual factors, such as user location, activity venue/time,
existing participants, and so on.
Mobile crowd sensing (MCS) [12] leverages crowdcontributed
data collected via smartphone sensors in the physical
space as well as mobile social networks (SNs) in the cyber
space. MCS has been employed in numerous application areas,
yet its use in group activity organization is underinvestigated.
Our work aims to exploit the cross-space sensing nature of MCS
to support the lifecycle of real-world group activities.
In this paper, we present the MobiGroup, a group-aware system
that provides assistance throughout various group activity
organizational stages. It exploits smartphone sensing to capture
online/offline social interactions and empowers group formation
and management. We extend [13] by 1) addressing the activity
lifecycle in real-world settings; 2) characterizing the complexity
and diversity of social activity organizational processes in a
formal concept model; and 3) providing intelligent facilitations
for social activity preparation. Specifically, our contributions
include the following.
1) A generic and multiviewed group activity model: The activity
model classifies the lifecycle of group activities into
four stages. Based on the model, we develop a framework
that can adapt supports to the characteristics and
organizational stages of a group activity.

2) Context-aware approaches to group activity preparation:
For activities that are open to the public, we propose a
heuristic rule-based strategy to disseminate information
of an activity according to its popularity and group preferences
[14]. For private activities that often consist of a
similar set of participants, we use a social graph model
to characterize the closed activity participation network
and develop a context-based group computing method for
highly relevant group recommendation.
3) Cross-community approach to ongoing activity suggestion:
To encourage participation of ongoing events, we
propose a mechanism for recommending activities to potentially
interested users by extracting static/dynamic interaction
features of both online and offline communities
[15].
The rest of the paper is organized as follows: Section II discusses
related work. Section III proposes the group activity concept
model, and Section IV presents the framework for group
activity organization. We describe our methods for planned activity
preparation and ongoing activity recommendation in SectionsVand
VI, respectively, and present an evaluation in Section
VII. We conclude the paper in Section VIII.