Anomaly Detection on Time Series Data

Abstract

Anomaly detection is an important problem that has been researched within diverse application domains. Detection of anomalies in the time series domain finds extensive application in monitoring system status, mal-ware/spam detection, credit-card fraud etc. In this work we explore methods to detect anomalies in multivariate as well as uni variate time-series and proposed a novel method using Dictionary Learning, Sparse Representation, Singular Value Decomposition and Topological anomaly detection(TAD). We have tested the proposed method on real as well as synthetic data sets. Our novel method brings down the false positive rates as compared to the existing methods.