A symbolic representation of time series, with implications for streaming algorithms

Lin, Jessica; Keogh, Eamonn; Lonardi, Stefano; Chiu, Bill

doi:10.1145/882082.882086

Cited by 1,179 publications

(669 citation statements)

References 33 publications

Supporting

Mentioning

599

Contrasting

Unclassified

Order By: Relevance

“…17 Therefore, it would follow that the down-regulation of this transcription factor would lead to the suppression of cellular proliferation, one of the hallmarks of malignant tumors. The down-regulation of USF is characteristic of the decrease in lipid and glucose metabolism by the liver, 23 leading to the increase in the levels of circulating free fatty acids and glucose in the bloodstream leading to the associated steroid induced diabetes.…”

Section: Discussionmentioning

confidence: 99%

Context Specific Transcription Factor Prediction

et al. 2007

View full text Add to dashboard Cite

Abstract-One of the goals of systems biology is the identification of regulatory mechanisms that govern an organism's response to external stimuli. Transcription factors have been hypothesized as a major contributor to an organism's response to various outside stimuli, and a great deal of work has been done to predict the set of transcription factors which regulate a given gene. Most of the current methods seek to identify possible binding sites from genomic sequence. Initial attempts at predicting transcription factors from genomic sequences suffered from the problem of false positives. Making the problem more difficult, it has also been shown that while predicted binding sites might be false positives, they can be shown to bind to their corresponding sequences in vitro. One method for rectifying this is through the use of phylogenetic analysis in which only regions which show high evolutionary conservation are analyzed. However such an approach may be too stringent because of the level of degeneracy shown in transcription factor binding site position weight matrices. Due to the degeneracy, there may be only a few bases that need to be conserved across species. Therefore, while a sequence may not show a high level of evolutionary conservation, these sequences may still show high affinity for the same transcription factor. In predicting transcription factor binding we explore the notion that ''Coexpression implies co-regulation'' [Allocco et al. BMC Bioinformatics 5:18, 2004]. With multiple genes requiring similar transcription factors binding sites, there exists a basis for eliminating false positives. This method allows for the selection of transcription factors binding sites that are active under a given experimental paradigm, thereby allowing us to indirectly incorporate the effects of chromosome and recognition site presentation upon transcription factor binding prediction. Rather than having to rationalize that a few transcription factors binding sites are over-represented in a cluster of genes, one can show that a few transcription factors are active in the cluster of genes that have been grouped together. Although the method focuses on predicting experiment-specific transcription factor binding sites, it is possible that if such a methodology were used in an iterative process where different experiments were analyzed, one could obtain a comprehensive set of transcription factors binding sites which regulate the various dynamic responses shown by biological systems under a variety of conditions hence building a more comprehensive model of transcriptional regulation.

show abstract

Section: Discussionmentioning

confidence: 99%

Context Specific Transcription Factor Prediction

et al. 2007

View full text Add to dashboard Cite

show abstract

“…The search for unknown motifs is at the heart of the work conducted by Keogh et al [18,19,20,9] . Given the emphasis on unknown patterns Keogh states "to the best of our knowledge, the problem of finding repeated patterns in time series has not been addressed (or even formulated) in the literature" [9] .…”

Section: Related Workmentioning

confidence: 99%

“…Whilst immunology provides the inspiration for the theory behind the MTA, the work of Keogh et al [20,9] is the inspiration for the initial implementation of the MTA. In particular, Keogh's SAX technique for representing a time series was a contributing factor in the success of the MTA.…”

Section: Motif Detection: Terms and Definitionsmentioning

confidence: 99%

See 1 more Smart Citation

The motif tracking algorithm

Wilson

Birkin

Aickelin

2008

Int. J. Autom. Comput.

View full text Add to dashboard Cite

Abstract:The search for patterns or motifs in data represents a problem area of key interest to finance and economic researchers. In this paper we introduce the Motif Tracking Algorithm, a novel immune inspired pattern identification tool that is able to identify unknown motifs of a non specified length which repeat within time series data. The power of the algorithm comes from the fact that it uses a small number of parameters with minimal assumptions regarding the data being examined or the underlying motifs. Our interest lies in applying the algorithm to financial time series data to identify unknown patterns that exist. The algorithm is tested using three separate data sets. Particular suitability to financial data is shown by applying it to oil price data. In all cases the algorithm identifies the presence of a motif population in a fast and efficient manner due to the utilisation of an intuitive symbolic representation. The resulting population of motifs is shown to have considerable potential value for other applications such as forecasting and algorithm seeding.

show abstract

“…Multivariate time series data classification methods were studied in [4,5,6,7,8], including On-demand Classifier [4], HMM (Hidden Markov Models) [5], RNN (Recurrent Neural Network), Dynamic Time Warping [5], weighted ensemble classifier [6] and SAX [7]. These methods involve large numbers of parameters and complex preprocessing step that need to be tuned.…”

Section: Related Workmentioning

confidence: 99%

Multivariate Stream Data Classification Using Simple Text Classifiers

Seo

Kang

Lee

et al. 2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. We introduce a classification framework for continuous multivariate stream data. The proposed approach works in two steps. In the preprocessing step, it takes as input a sliding window of multivariate stream data and discretizes the data in the window into a string of symbols that characterize the signal changes. In the classification step, it uses a simple text classification algorithm to classify the discretized data in the window. We evaluated both supervised and unsupervised classification algorithms. For supervised, we tested Naïve Bayes Model and SVM, and for unsupervised, we tested Jaccard, TFIDF, Jaro and JaroWinkler. In our experiments, SVM and TFIDF outperformed the other classification methods. In particular, we observed that classification accuracy is improved when the correlation of attributes is also considered along with the n-gram tokens of symbols.

show abstract

A symbolic representation of time series, with implications for streaming algorithms

Cited by 1,179 publications

References 33 publications

Context Specific Transcription Factor Prediction

Context Specific Transcription Factor Prediction

The motif tracking algorithm

Multivariate Stream Data Classification Using Simple Text Classifiers

Contact Info

Product

Resources

About