Locally adaptive dimensionality reduction for indexing large time series databases

Keogh, Eamonn; Chakrabarti, Kaushik; Pazzani, Michael J.; Mehrotra, Sharad

doi:10.1145/375663.375680

Cited by 525 publications

(304 citation statements)

References 41 publications

Supporting

Mentioning

299

Contrasting

Unclassified

Order By: Relevance

“…In addition to its usefulness in the trajectory database, the similarity query is one of the most interesting fields in time series databases. In time series databases, the similarity between two sets of time series data is typically measured by the Euclidean distance [6] [7], which can be calculated efficiently. However, there have been few discussions on the similarity between two lines in space because the previous approaches for spatial queries have focused on the "distance" between a point and a line [2] [9] [15].…”

Section: Shape-based Approachmentioning

confidence: 99%

“…For the time series database, the similarity of the two time series data, where each has n value, is given by the Euclidean distance between vectors in R n . In [6] and [7], when there are two time series data, c = w 1 , w 2 , . .…”

Section: Shape-based Similarity Querymentioning

confidence: 99%

“…However, we found this useful idea through research of time series databases [6] [7] [10]. The time series database systems can store time series data such as temperature, economic indicators, population, wave signals, and so on, in addition to supporting queries for extracting patterns from the time series data.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Shape-Based Similarity Query for Trajectory of Mobile Objects

Yanagisawa¹,

Akahani²,

Satoh³

2002

Lecture Notes in Computer Science

113

View full text Add to dashboard Cite

Abstract. In this paper, we describe an efficient indexing method for a shape-based similarity search of the trajectory of dynamically changing locations of people and mobile objects. In order to manage trajectories in database systems, we define a data model of trajectories as directed lines in a space, and the similarity between trajectories is defined as the Euclidean distance between directed discrete lines. Our proposed similarity query can be used to find interested patterns embedded into the trajectories, for example, the trajectories of mobile cars in a city may include patterns for expecting traffic jams. Furthermore, we propose an efficient indexing method to retrieve similar trajectories for a query by combining a spatial indexing technique (R + -Tree) and a dimension reduction technique, which is called PAA (Piecewise Approximate Aggregate). The indexing method can efficiently retrieve trajectories whose shape in a space is similar to the shape of a candidate trajectory from the database.

show abstract

Section: Shape-based Approachmentioning

confidence: 99%

Section: Shape-based Similarity Querymentioning

confidence: 99%

See 1 more Smart Citation

Shape-Based Similarity Query for Trajectory of Mobile Objects

Yanagisawa¹,

Akahani²,

Satoh³

2002

Lecture Notes in Computer Science

113

View full text Add to dashboard Cite

show abstract

“…The proposed methods mainly differ in the representation of the time series, a survey is given in [1]. Standard techniques for dimension reduction include Discrete Fourier Transform (e.g.…”

Section: Introductionmentioning

confidence: 99%

TQuEST: Threshold Query Execution for Large Sets of Time Series

Aßfalg¹,

Kriegel²,

Kröger³

et al. 2006

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. Effective and efficient data mining in time series databases is essential in many application domains as for instance in financial analysis, medicine, meteorology, and environmental observation. In particular, temporal dependencies between time series are of capital importance for these applications. In this paper, we present TQuEST, a powerful query processor for time series databases. TQuEST supports a novel but very useful class of queries which we call threshold queries. Threshold queries enable searches for time series whose values are above a user defined threshold at certain time intervals. Example queries are "report all ozone curves which are above their daily mean value at the same time as a given temperature curve exceeds 28 • C" or "report all blood value curves from patients whose values exceed a certain threshold one hour after the new medication was taken". TQuEST is based on a novel representation of time series which allows the query processor to access only the relevant parts of the time series. This enables an efficient execution of threshold queries. In particular, queries can be readjusted with interactive response times.

show abstract

“…Several high level dimensionality reductions of time series have been proposed, including Singular Value Decomposition (SVD) [7], the Discrete Fourier transform (DFT) [8], Discrete Wavelets Transform (DWT) [9], Symbolic Mappings [10,11,12], Piecewise Linear Representation (PLR) [13] and Piecewise Aggregate Approximation (PAA) [14]. As a matter of fact, SVD, DFT and DWT are mainly designed to match patterns, so they are relatively intricate to be implemented to compress data.…”

Section: Compression Algorithmsmentioning

confidence: 99%

Efficient Time Series Data Classification and Compression in Distributed Monitoring

Jin

et al.

Emerging Technologies in Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

Abstract. As a key issue in distributed monitoring, time series data are a series of values collected in terms of sequential time stamps. Requesting them is one of the most frequent requests in a distributed monitoring system. However, the large scale of these data users request may not only cause heavy loads to the clients, but also cost long transmission time. In order to solve the problem, we design an efficient two-step method: first classify various sets of time series according to their sizes, and then compress the time series with relatively large size by appropriate compression algorithms. Finally, we test our method, and the results show that: 1) This two-step approach is really able to reduce the users' response time after requesting the monitoring data.2) The compression effects of the algorithms designed by us are satisfactory.

show abstract

Locally adaptive dimensionality reduction for indexing large time series databases

Abstract: Abstract

Cited by 525 publications

References 41 publications

Shape-Based Similarity Query for Trajectory of Mobile Objects

Shape-Based Similarity Query for Trajectory of Mobile Objects

TQuEST: Threshold Query Execution for Large Sets of Time Series

Efficient Time Series Data Classification and Compression in Distributed Monitoring

Contact Info

Product

Resources

About