Time series representation for anomaly detection

Leng, Mingwei; Lai, Xinsheng; Guo-lv, Tan; Xu, Xia

doi:10.1109/iccsit.2009.5234775

Cited by 16 publications

(5 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Thatte et al used packet-sized statistics and traffic rates to build a statistical model and then employed the sequential probability ratio test (SPRT) in the detection phase [14]. Some other methods in this category work by considering normal behavior as a time series and then combining it with signal processing techniques for anomaly detection purposes [15]. Hajji [10] stated that normal data are considered to be generated by a set of normal distributions, and all parameters of the model are generated based on the existent data; any deviation from these models is considered an anomaly.…”

Section: Methodsmentioning

confidence: 98%

A fast and noise resilient cluster-based anomaly detection

Bigdeli

Mohammadi

Raahemi

et al. 2015

Pattern Anal Applic

View full text Add to dashboard Cite

Clustering, while systematically applied in anomaly detection, has a direct impact on the accuracy of the detection methods. Existing cluster-based anomaly detection methods are mainly based on spherical shape clustering. In this paper, we focus on arbitrary shape clustering methods to increase the accuracy of the anomaly detection. However, since the main drawback of arbitrary shape clustering is its high memory complexity, we propose to summarize clusters first. For this, we design an algorithm, called Summarization based on Gaussian Mixture Model (SGMM), to summarize clusters and represent them as Gaussian Mixture Models (GMMs). After GMMs are constructed, incoming new samples are presented to the GMMs, and their membership values are calculated, based on which the new samples are labeled as ''normal'' or ''anomaly.'' Additionally, to address the issue of noise in the data, instead of labeling samples individually, they are clustered first, and then each cluster is labeled collectively. For this, we present a new approach, called Collective Probabilistic Anomaly Detection (CPAD), in which, the distance of the incoming new samples and the existing SGMMs is calculated, and then the new cluster is labeled the same as of the closest cluster. To measure the distance of two GMM-based clusters, we propose a modified version of the Kullback-Libner measure. We run several experiments to evaluate the performances of the proposed SGMM and CPAD methods and compare them against some of the wellknown algorithms including ABACUS, local outlier factor (LOF), and one-class support vector machine (SVM). The performance of SGMM is compared with ABACUS using Dunn and DB metrics, and the results indicate that the SGMM performs superior in terms of summarizing clusters. Moreover, the proposed CPAD method is compared with the LOF and one-class SVM considering the performance criteria of (a) false alarm rate, (b) detection rate, and (c) memory efficiency. The experimental results show that the CPAD method is noise resilient, memory efficient, and its accuracy is higher than the other methods.

show abstract

Section: Methodsmentioning

confidence: 98%

A fast and noise resilient cluster-based anomaly detection

Bigdeli

Mohammadi

Raahemi

et al. 2015

Pattern Anal Applic

View full text Add to dashboard Cite

show abstract

“…It also explores events that have surprising effects on the process of time series (Keogh et al, 2002), (Leng et al, 2009). Clustering categorizes data by reducing the volume of data and finding patterns.…”

Section: Literature Reviewmentioning

confidence: 99%

A Comparative Approach to Financial Clustering Models: (A Study of the Companies Listed on Tehran Stock Exchange)

Nourahmadi¹,

Rasti²,

Sadeqi³

2022

IJFIFSA

View full text Add to dashboard Cite

show abstract

“…Applications of time series clustering can be divided into several types of problems: anomaly or discord detection -finding unexpected and/or unwanted patterns in the series Chan & Mahoney, 2005;Wei et al, 2005;Leng et al, 2009); recognizing dynamic changes in the time series -correlation analysis (He et al, 2011); prediction and recommendation -various techniques of clustering and approximation enabling prediction of the future behavior of time series (Sfetsos & Siriopoulos, 2004;Pavlidis et al, 2006;Graves & Pedrycz, 2010); and pattern recognition -finding interesting patterns in databases, behavior patterns of sales or trading data (Kumar & Patel, 2002;Guan & Jiang, 2007;Aghabozorgi & Wah, 2014).…”

Section: Introductionmentioning

confidence: 99%

Hierarchical clustering of time series data with parametric derivative dynamic time warping

Łuczak

2016

Expert Systems with Applications

View full text Add to dashboard Cite

Time series representation for anomaly detection

Cited by 16 publications

References 6 publications

A fast and noise resilient cluster-based anomaly detection

A fast and noise resilient cluster-based anomaly detection

A Comparative Approach to Financial Clustering Models: (A Study of the Companies Listed on Tehran Stock Exchange)

Hierarchical clustering of time series data with parametric derivative dynamic time warping

Contact Info

Product

Resources

About