Alternative Quality Measures for Time Series Shapelets

Lines, Jason; Bagnall, Anthony

doi:10.1007/978-3-642-32639-4_58

Cited by 27 publications

(10 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Distances between series and shapelets represent shapelet-transformed [10] classification features for a series of segregation metrics, such as information gain [17,11], FStat [8] or Kruskall-Wallis [9]. The brute-force candidates search approach, based on an exhaustive search of candidates, suffers from Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page.…”

Section: Introductionmentioning

confidence: 44%

See 1 more Smart Citation

Learning time-series shapelets

Grabocka

Schilling

Wistuba

et al. 2014

Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

358

301

View full text Add to dashboard Cite

Shapelets are discriminative sub-sequences of time series that best predict the target variable. For this reason, shapelet discovery has recently attracted considerable interest within the time-series research community. Currently shapelets are found by evaluating the prediction qualities of numerous candidates extracted from the series segments. In contrast to the state-of-the-art, this paper proposes a novel perspective in terms of learning shapelets. A new mathematical formalization of the task via a classification objective function is proposed and a tailored stochastic gradient learning algorithm is applied. The proposed method enables learning nearto-optimal shapelets directly without the need to try out lots of candidates. Furthermore, our method can learn true top-K shapelets by capturing their interaction. Extensive experimentation demonstrates statistically significant improvement in terms of wins and ranks against 13 baselines over 28 time-series datasets.

show abstract

Section: Introductionmentioning

confidence: 44%

“…Apart from their high prediction accuracy, shapelets also offer interpretable features to domain experts. Moreover, discovering shapelets has been a hot topic in the time-series domain during the last five years [17,11,10,19,8,12,9].…”

Section: Introductionmentioning

confidence: 46%

Learning time-series shapelets

Grabocka

Schilling

Wistuba

et al. 2014

Proceedings of the 20th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

358

301

View full text Add to dashboard Cite

show abstract

“…Quality can be assessed by information gain [17] or alternative measures such as the F, moods median or rank order statistic [11]. Once all the shapelets for a series are evaluated they are sorted and the lowest quality overlapping shapelets are removed.…”

Section: Shapelet Based Classificationsupporting

confidence: 45%

Binary Shapelet Transform for Multiclass Time Series Classification

Bostrom

Bagnall

2015

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Abstract. Shapelets have recently been proposed as a new primitive for time series classification. Shapelets are subseries of series that best split the data into its classes. In the original research, shapelets were found recursively within a decision tree through enumeration of the search space. Subsequent research indicated that using shapelets as the basis for transforming datasets leads to more accurate classifiers. Both these approaches evaluate how well a shapelet splits all the classes. However, often a shapelet is most useful in distinguishing between members of the class of the series it was drawn from against all others. To assess this conjecture, we evaluate a one vs all encoding scheme. This technique simplifies the quality assessment calculations, speeds up the execution through facilitating more frequent early abandon and increases accuracy for multi-class problems. We also propose an alternative shapelet evaluation scheme which we demonstrate significantly speeds up the full search.

show abstract

“…To find the split point and measure the relevance in a one dimensional orderline (as in Figure 5) have been proposed several techniques, such as the maximum information gain [6], the Kruskal-Wallis and Mood's Median [5], and the Left Side Pure (LSP) [3]. The LSP returns a split point between T 1 and T 5 in order to keep the left side of the orderline pure.…”

Section: Relevance Measuring For Multidimensional Subtrajectory Candimentioning

confidence: 99%