Device Graphing by Example

Funkhouser, Keith; Malloy, Matthew; Ceyhun, Enis; Poon, Phillip K.; Barford, Paul

doi:10.1145/3219819.3219852

Cited by 5 publications

(10 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Methodologies presented by (Malloy et al 2017;Funkhouser et al 2018) were built on the concept of IP colocation, by observing co-occurrences of IDs from an IP address at a specific point in time. IP co-location graphs formed the foundation of their device graphs.…”

Section: Related Workmentioning

confidence: 99%

“…Classification models were used to predict relationships between devices. Beyond IP and device-related features (Volkova 2017), additional learning features could be gleaned from browsing logs and associated meta-data (Tran 2016;Funkhouser et al 2018) that may be useful. To improve the classifier, (Tran 2016) proposed several features such as domains visited, actions taken, and time spent while browsing.…”

Section: Related Workmentioning

confidence: 99%

“…One of the biggest challenges of developing device graphs is accurately validating the user-device linkages. In this regard, (Funkhouser et al 2018) used data from user panels operated by Comscore as a source of ground truth, realized through specialized monitoring software that provided data such as unique hardware identifiers of devices and associated IDs. Such data was unavailable in our context and infeasible for commercial implementation due to the amount of third-party involvement required.…”

Section: Related Workmentioning

confidence: 99%

“…From the literature, there also appeared to be limited testing of generated graphs. For example, in (Malloy et al 2017;Funkhouser et al 2018), graphs were built using data from Comscore's digital census network consisting of 700 million records collected over six weeks in the United States, but were not evaluated regularly over a period of time. Graphs appeared to be one-off output and were not updated continuously to capture new identifiers.…”

Section: Related Workmentioning

confidence: 99%

“…First-party data gathered across major properties owned by Mediacorp was used in the exercise. We referenced existing methodologies (Malloy et al 2017;Funkhouser et al 2018) that utilized IP-related features to generate the device graph. We implemented a new layer in the graph construction process by generating additional behavioral features (e.g.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Cross-Device Identity Resolution using Machine Learning: A Scalable Device Graph Approach

Koduri¹,

Lim

Li³

et al. 2021

FLAIRS

View full text Add to dashboard Cite

This paper outlines how Mediacorp, Singapore's public service broadcaster, addresses its cross-device identity challenges using a scalable device graph approach. Research in this area is relevant to the domain of advertising technology as it enables a holistic view of consumers that can be extended to use cases such as improving advertisement targeting, personalized recommendations and demographic predictions. Past research efforts were limited to high-level descriptions of the steps undertaken to create a one-off, static device graph based on data collected over a circumscribed time frame, thus limiting its use in larger-scale commercial applications. In this paper, we propose a scalable solution that enables continuous, incremental revisions of our device graph. We leverage behavioral data captured by Mediacorp across its sites and platforms to build a richer device graph that is updated weekly. First, we introduce additional features and explore various classifiers to improve pairwise probability scores between devices that are likely to belong to the same user. Then, we apply community clustering algorithms to uncover device communities to establish the final device graph. Extensive experiments showed that our additive approach has consistently delivered >90% precision and recall in real-world applications.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations