Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Fast and accurate time-series clustering
Paparrizos J., Gravano L. ACM Transactions on Database Systems42 (2):1-49,2017.Type:Article
Date Reviewed: Apr 16 2018

Clustering temporal data, namely time series, is a challenging and expensive computational task in terms of accuracy and speed. Despite the fact that a wide variety of time-series clustering algorithms exist in the literature, they remain unsatisfactory because their computational performance is doomed with a bottleneck: when good accuracy is achieved, then execution is slow, and vice versa. Remarkably, one algorithm (for example, k-medoids) delivers very good performance on both endeavors (that is, steadfast accuracy and fast execution speed). However, clustering, in its nature, should be an unsupervised learning algorithm; yet, k-medoids requires supervision (that is, tuning of its parameters by an expert).

In this regard, the authors provide a clustering algorithm for time series that is accurate, fast, and parameter free. Their work is primarily based on the shape base distance (SBD) measure, which is commonly used to compare temporal data. Then, based on SBD, they develop two novel methods for time-series analysis: ShapeExtraction (SE) and MultiShapesExtraction (MSE). Subsequently, the authors develop temporal data clustering using SE and MSE. The first clustering algorithm is called k-shape and uses the SE method to find the center of a cluster based on all temporal data of each cluster. The second clustering algorithm is called k-MultiShape (k-MS) and uses MSE to find multiple centers per cluster and compare them to the time-series temporal data of each cluster.

Exhaustive theoretical and experimental proofs are provided in this paper, which make it a valuable reference for advanced research fellows tackling time-series analysis and clustering. In addition, the intense literature offered and meticulous comparative study (to other algorithms) afforded, accompanied with interesting and outstanding results, fulfill all lacunas that I could come up with, and erased, on the go, all negative criticisms that I was thinking of while reading the paper. Thus, I can only encourage you, the reader who reached the end of this review, to dig into this paper; I anticipate its fruitful application in your domain of interest.

Reviewer:  Mario Antoine Aoun Review #: CR145977 (1806-0330)
Bookmark and Share
  Reviewer Selected
Featured Reviewer
 
 
Clustering (H.3.3 ... )
 
 
Data Mining (H.2.8 ... )
 
 
Time Series Analysis (G.3 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Clustering": Date
Concepts and effectiveness of the cover-coefficient-based clustering methodology for text databases
Can F. (ed), Ozkarahan E. ACM Transactions on Database Systems 15(3): 483-517, 1990. Type: Article
Dec 1 1992
A parallel algorithm for record clustering
Omiecinski E., Scheuermann P. ACM Transactions on Database Systems 15(3): 599-624, 1990. Type: Article
Nov 1 1992
Organization of clustered files for consecutive retrieval
Deogun J., Raghavan V., Tsou T. ACM Transactions on Database Systems 9(4): 646-671, 1984. Type: Article
Jun 1 1985
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy