VEDAR: Accountable Behavioural Change Detection
With exponential increase in the availability oftelemetry / streaming / real-time data, understanding contextualbehavior changes is a vital functionality in order to deliverunrivalled customer experience and build high performance andhigh availability systems. Real-time behavior change detectionfind...
Gespeichert in:
Hauptverfasser: | , , , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | With exponential increase in the availability oftelemetry / streaming /
real-time data, understanding contextualbehavior changes is a vital
functionality in order to deliverunrivalled customer experience and build high
performance andhigh availability systems. Real-time behavior change
detectionfinds a use case in number of domains such as social networks,network
traffic monitoring, ad exchange metrics etc. In streamingdata, behavior change
is an implausible observation that does notfit in with the distribution of rest
of the data. A timely and preciserevelation of such behavior changes can give
us substantialinformation about the system in critical situations which can bea
driving factor for vital decisions. Detecting behavior changes instreaming
fashion is a difficult task as the system needs to processhigh speed real-time
data and continuously learn from data alongwith detecting anomalies in a single
pass of data. In this paperwe introduce a novel algorithm called Accountable
BehaviorChange Detection (VEDAR) which can detect and elucidate thebehavior
changes in real-time and operates in a fashion similarto human perception. We
have bench marked our algorithmon open source anomaly detection datasets. We
have benchmarked our algorithm by comparing its performance on opensource
anomaly datasets against industry standard algorithmslike Numenta HTM and
Twitter AdVec (SH-ESD). Our algorithmoutperforms above mentioned algorithms for
behaviour changedetection, efficacy is given in section V. |
---|---|
DOI: | 10.48550/arxiv.1902.06663 |