Princeton University Library Catalog

Clustering and Outlier Detection: Methods and Applications in Smart Home Networks

Author/​Artist:
Davis, Erick [Browse]
Format:
Senior thesis
Language:
English
Advisor(s):
Kpotufe, Samory [Browse]
Department:
Princeton University. Department of Operations Research and Financial Engineering [Browse]
Class year:
2016
Description:
96 pages
Summary note:
We present a review of outlier detection techniques and attempt to find a consensus approach for defining an outlier. We then extensively explore device activity in smart home networks. We attempt to identify individual device network profiles and differentiate the behavior originating from device activities. This is formulated as a clustering problem. Device behavior is analyzed in the context of finding an appropriate outlier detection definition and approach.