Efficient Learning Interpretable Shapelets for Accurate Time Series Classification

Fang, Zicheng; Wang, Peng; Wang, Wei

doi:10.1109/icde.2018.00052

Cited by 38 publications

(29 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Content may change prior to final publication. return wordLength 15: end if each index of the window can be computed by (16) to form F (Line 3). After that, p 1 and p 2 are computed according to (17) and (18) (Lines 4-5), respectively.…”

Section: A Learning the Word Length For Each Sliding Windowmentioning

confidence: 99%

See 1 more Smart Citation

Discriminative Dictionary Learning for Time Series Classification

et al. 2020

View full text Add to dashboard Cite

Time series symbolization based on the Symbolic Fourier Approximation (SFA) and a sliding window mechanism can effectively improve classification performance. Hence, it has become a research hotspot of time series representation learning. However, there are still some obvious shortcomings. First, the lengths of the generated words are consistent for different sliding windows in the symbolization process, ignoring the difference of the discriminative information contained in diverse periods. Second, research on the relationship between words is limited to adjacent ones. Third, existing dictionary learning methods do not consider a filtering algorithm for discriminative word selection. To this end, a novel and fast variable-length word generation method, incorporating the skip-bigram model for the construction of symbiotic word pairs, is proposed for the first time in our work. Then, a discriminative word filter with a dynamic threshold is designed to build the discriminative word dictionary. Many controlled experiments first verified the effectiveness of each proposed method. Then, the performance improvement ability of the word dictionary is demonstrated by the comparative experiments with the representative classification models based on different theoretical foundations.

show abstract

Section: A Learning the Word Length For Each Sliding Windowmentioning

confidence: 99%

“…Hence, local feature-based models first need to reconstruct the features of time series. After defining the feature prototype, one type trains the classification model directly [13]- [16] while the other first tries to transform the time series data, and then apply different classification models on the transformed data [17]- [23].…”

Section: Introductionmentioning

confidence: 99%

Discriminative Dictionary Learning for Time Series Classification

et al. 2020

View full text Add to dashboard Cite

show abstract

“…One type of the method utilizes the top- k shapelets to create a transformed dataset, on which the traditional classification algorithms [ 20 – 24 ] could be applied. The other uses the shapelets to build the classification model directly [ 18 , 25 – 28 ].…”

Section: Introductionmentioning

confidence: 99%

“…Rakthanmanon et al [ 26 ] proposed a fast shapelet discovery algorithm based on Symbolic Aggregate approXimation (SAX). Similarly, Fang et al [ 28 ] introduced a novel method to search shapelets based on piecewise aggregate approximation (PAA). (3) A random selection mechanism is used to select shapelets.…”

Section: Introductionmentioning

confidence: 99%

Shapelet Discovery by Lazy Time Series Classification

Zhang

Wang

Yuan

et al. 2020

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

As a representation of discriminative features, the time series shapelet has recently received considerable research interest. However, most shapelet-based classification models evaluate the differential ability of the shapelet on the whole training dataset, neglecting characteristic information contained in each instance to be classified and the classwise feature frequency information. Hence, the computational complexity of feature extraction is high, and the interpretability is inadequate. To this end, the efficiency of shapelet discovery is improved through a lazy strategy fusing global and local similarities. In the prediction process, the strategy learns a specific evaluation dataset for each instance, and then the captured characteristics are directly used to progressively reduce the uncertainty of the predicted class label. Moreover, a shapelet coverage score is defined to calculate the discriminability of each time stamp for different classes. The experimental results show that the proposed method is competitive with the benchmark methods and provides insight into the discriminative features of each time series and each type in the data.

show abstract

“…As for the scenario that the occurrence of specific sub-series determines a class, TS can then be represented by such shape-based features, namely Shapelet [14]. Various Shapelet-based approaches have been proposed to optimize both the accuracy [15,16] and the efficiency [17,18] of the classification. Another remarkable attempt [19][20][21] adopting ensemble approaches on several TS representations (e.g., Shapelet-based, similarity-based, interval-based etc.)…”

Section: Introductionmentioning

confidence: 99%

Incremental and Adaptive Feature Exploration over Time Series Stream

Zuo

Zeitouni

Taher

2019

2019 IEEE International Conference on Big Data (Big Data)

View full text Add to dashboard Cite

Over past years, various attempts have been made at analysing Time Series (TS) which has been raising great interest of Data Mining community due to its special data format and broad application scenarios. An important aspect in TS analysis is Time Series Classification (TSC), which has been applied in medical diagnosis, human activity recognition, industrial troubleshooting, etc. Typically, all TSC work trains a stable model from an off-line TS dataset, without considering potential Concept Drift in streaming context. Domains like healthcare look to enrich the database gradually with more medical cases, or in astronomy, with human's growing knowledge about the universe, the theoretical basis for labelling data will change. The techniques applied in a stable TS dataset are then not adaptable in such dynamic scenarios (i.e. streaming context). Classical data stream analysis are biased towards vector or row data, where each attribute is independent to train an adaptive learning model, but rarely considers Time Series as a stream instance. Processing such type of data requires combining techniques in both communities of Time Series (TS) and Data Streams. To this end, by adopting the concepts of Shapelet and Matrix Profile, we conduct the first attempt to extract the adaptive features from Time Series Stream based on the Test-then-Train strategy, which is applicable in both contexts: a) under stable concept, learning model will be updated incrementally; b) for data source with Concept Drift, previous concepts that do not represent the current stream behavior will be discarded from the model.

show abstract

Efficient Learning Interpretable Shapelets for Accurate Time Series Classification

Cited by 38 publications

References 32 publications

Discriminative Dictionary Learning for Time Series Classification

Discriminative Dictionary Learning for Time Series Classification

Shapelet Discovery by Lazy Time Series Classification

Incremental and Adaptive Feature Exploration over Time Series Stream

Contact Info

Product

Resources

About