Presence analytics: Detecting classroom-based social patterns using WLAN traces

Muawya Habib Sarnoub Eldaw, Mark Levene, George Roussos

Research output: Contribution to Book/ReportConference Contributionpeer-review

Abstract

We demonstrate how density-based clustering of WLAN traces can be utilised to discover social groups of students within a university campus. For this purpose we deploy a temporally restricted version of the Social-DBSCAN algorithm [2] to discover social groups of students who attend the same classes. We detect the existence of social relationships between the students attending the same class by analysing their behaviour of visit during break-times. The intuition is that if a group of two or more students are friends, who attend the same classes, then they are likely to be socialising/meeting more often at locations such as the Coffee-shop during break-times. By leveraging information extracted from the timetable as well as the teaching practices at the case-study university, we inform our model about the duration of the break-times. Utilising a large data set of Eduroam traces, collected at the main site of the case-study institution, we chose as a proof concept, a set of locations for the evaluation of the proposed method, which we successfully employed to detect the social groups of students who attended regular classes at those chosen locations.
Original languageEnglish
Title of host publication2017 Intelligent Systems Conference (IntelliSys)
PublisherIEEE
ISBN (Print)9781509064359
DOIs
Publication statusPublished - 30 Sept 2017
Externally publishedYes
EventIntelligent Systems Conference 2017 - London, United Kingdom
Duration: 7 Sept 20178 Sept 2017

Conference

ConferenceIntelligent Systems Conference 2017
Country/TerritoryUnited Kingdom
CityLondon
Period7/09/178/09/17

Fingerprint

Dive into the research topics of 'Presence analytics: Detecting classroom-based social patterns using WLAN traces'. Together they form a unique fingerprint.

Cite this