Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

W4-Groups: Modeling the Who, What, When and Where of Group Behavior via Mobility Sensing

Akanksha Atrey, Camellia Zakaria, Rajesh Balan, Prashant Shenoy

TL;DR

W4-Groups introduces a mobility-sensing framework to model who-what-when-where in group interactions, capturing short-term and long-term formations across multiple data sources. It combines sessions, pairwise co-occurrences, and long-term mobility similarity (spatial, temporal, social) to detect dynamic groups and filter false positives, achieving around 92–93% accuracy across WiFi and check-in datasets. The approach enables applications such as next-activity prediction and longitudinal analyses of behavior changes (e.g., pre/post COVID-19), and demonstrates robustness and generalizability via cross-dataset evaluation. The work also discusses privacy safeguards and ethical considerations for deployment in real-world contexts.

Abstract

Human social interactions occur in group settings of varying sizes and locations, depending on the type of social activity. The ability to distinguish group formations based on their purposes transforms how group detection mechanisms function. Not only should such tools support the effective detection of serendipitous encounters, but they can derive categories of relation types among users. Determining who is involved, what activity is performed, and when and where the activity occurs are critical to understanding group processes in greater depth, including supporting goal-oriented applications (e.g., performance, productivity, and mental health) that require sensing social factors. In this work, we propose W4-Groups that captures the functional perspective of variability and repeatability when automatically constructing short-term and long-term groups via multiple data sources (e.g., WiFi and location check-in data). We design and implement W4-Groups to detect and extract all four group features who-what-when-where from the user's daily mobility patterns. We empirically evaluate the framework using two real-world WiFi datasets and a location check-in dataset, yielding an average of 92% overall accuracy, 96% precision, and 94% recall. Further, we supplement two case studies to demonstrate the application of W4-Groups for next-group activity prediction and analyzing changes in group behavior at a longitudinal scale, exemplifying short-term and long-term occurrences.

W4-Groups: Modeling the Who, What, When and Where of Group Behavior via Mobility Sensing

TL;DR

W4-Groups introduces a mobility-sensing framework to model who-what-when-where in group interactions, capturing short-term and long-term formations across multiple data sources. It combines sessions, pairwise co-occurrences, and long-term mobility similarity (spatial, temporal, social) to detect dynamic groups and filter false positives, achieving around 92–93% accuracy across WiFi and check-in datasets. The approach enables applications such as next-activity prediction and longitudinal analyses of behavior changes (e.g., pre/post COVID-19), and demonstrates robustness and generalizability via cross-dataset evaluation. The work also discusses privacy safeguards and ethical considerations for deployment in real-world contexts.

Abstract

Human social interactions occur in group settings of varying sizes and locations, depending on the type of social activity. The ability to distinguish group formations based on their purposes transforms how group detection mechanisms function. Not only should such tools support the effective detection of serendipitous encounters, but they can derive categories of relation types among users. Determining who is involved, what activity is performed, and when and where the activity occurs are critical to understanding group processes in greater depth, including supporting goal-oriented applications (e.g., performance, productivity, and mental health) that require sensing social factors. In this work, we propose W4-Groups that captures the functional perspective of variability and repeatability when automatically constructing short-term and long-term groups via multiple data sources (e.g., WiFi and location check-in data). We design and implement W4-Groups to detect and extract all four group features who-what-when-where from the user's daily mobility patterns. We empirically evaluate the framework using two real-world WiFi datasets and a location check-in dataset, yielding an average of 92% overall accuracy, 96% precision, and 94% recall. Further, we supplement two case studies to demonstrate the application of W4-Groups for next-group activity prediction and analyzing changes in group behavior at a longitudinal scale, exemplifying short-term and long-term occurrences.
Paper Structure (50 sections, 13 equations, 9 figures, 6 tables, 1 algorithm)

This paper contains 50 sections, 13 equations, 9 figures, 6 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Motivating W4-Groups
  3. Supporting Group Modeling Applications
  4. Defining a Group
  5. Design Rationale
  6. W4-Groups System Design
  7. System Overview
  8. Data Processing Module
  9. Group Detection Module
  10. Modeling Sessions
  11. Modeling Pairwise Co-Occurrence
  12. Modeling Long-Term Mobility Behavior and Feature Extraction
  13. Long-Term Mobility Similarity Detection
  14. Detecting W4-Groups
  15. Prototype Implementation
  16. ...and 35 more sections

Figures (9)

  • Figure 1: From left to right, network trace of user mobility can produce three key attributes who-what-when-where in the group, reflecting the changes in user location throughout the day. We can infer the pairwise interaction of users collocating by processing these traces collectively at each place over time. Multiple collocations of the same users increase their certainty of existing as a group rather than collocating by chance.
  • Figure 2: W4-Groups overview consisting of two modules: data processing and group detection.
  • Figure 3: The inner workings of W4-Groups's modeling approach to produce who-what-when-where as its output.
  • Figure 4: Results of analyzing system parameters and performance tradeoffs: (a) impact of the length of long-term data; (b) impact of the similarity metric weights where 'spa' refers to spatial, 'tem' refers to temporal, and 'soc' refers to social; and (c) performance across group activity types.
  • Figure 4: Generalized performance across datasets.
  • ...and 4 more figures

Theorems & Definitions (8)

  • Definition 1
  • Definition 2
  • Definition 3
  • Definition 4
  • Definition 5
  • Definition 6
  • Definition 7
  • Definition 8