Generalised interpretable shapelets for irregular time series
The shapelet transform is a form of feature extraction for time series, in which a time series is described by its similarity to each of a collection of `shapelets'. However it has previously suffered from a number of limitations, such as being limited to regularly-spaced fully-observed time se...
| Main Authors: | , , |
|---|---|
| Format: | Internet publication |
| Language: | English |
| Published: |
2020
|
| _version_ | 1797109860938022912 |
|---|---|
| author | Kidger, P James, M Lyons, T |
| author_facet | Kidger, P James, M Lyons, T |
| author_sort | Kidger, P |
| collection | OXFORD |
| description | The shapelet transform is a form of feature extraction for time series, in which a time series is described by its similarity to each of a collection of `shapelets'. However it has previously suffered from a number of limitations, such as being limited to regularly-spaced fully-observed time series, and having to choose between efficient training and interpretability. Here, we extend the method to continuous time, and in doing so handle the general case of irregularly-sampled partially-observed multivariate time series. Furthermore, we show that a simple regularisation penalty may be used to train efficiently without sacrificing interpretability. The continuous-time formulation additionally allows for learning the length of each shapelet (previously a discrete object) in a differentiable manner. Finally, we demonstrate that the measure of similarity between time series may be generalised to a learnt pseudometric. We validate our method by demonstrating its performance and interpretability on several datasets; for example we discover (purely from data) that the digits 5 and 6 may be distinguished by the chirality of their bottom loop, and that a kind of spectral gap exists in spoken audio classification. |
| first_indexed | 2024-03-07T07:47:14Z |
| format | Internet publication |
| id | oxford-uuid:9c4a0b05-5a23-4816-a1bb-9367c0ac0753 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-07T07:47:14Z |
| publishDate | 2020 |
| record_format | dspace |
| spelling | oxford-uuid:9c4a0b05-5a23-4816-a1bb-9367c0ac07532023-06-09T12:53:41ZGeneralised interpretable shapelets for irregular time seriesInternet publicationhttp://purl.org/coar/resource_type/c_7ad9uuid:9c4a0b05-5a23-4816-a1bb-9367c0ac0753EnglishSymplectic Elements2020Kidger, PJames, MLyons, TThe shapelet transform is a form of feature extraction for time series, in which a time series is described by its similarity to each of a collection of `shapelets'. However it has previously suffered from a number of limitations, such as being limited to regularly-spaced fully-observed time series, and having to choose between efficient training and interpretability. Here, we extend the method to continuous time, and in doing so handle the general case of irregularly-sampled partially-observed multivariate time series. Furthermore, we show that a simple regularisation penalty may be used to train efficiently without sacrificing interpretability. The continuous-time formulation additionally allows for learning the length of each shapelet (previously a discrete object) in a differentiable manner. Finally, we demonstrate that the measure of similarity between time series may be generalised to a learnt pseudometric. We validate our method by demonstrating its performance and interpretability on several datasets; for example we discover (purely from data) that the digits 5 and 6 may be distinguished by the chirality of their bottom loop, and that a kind of spectral gap exists in spoken audio classification. |
| spellingShingle | Kidger, P James, M Lyons, T Generalised interpretable shapelets for irregular time series |
| title | Generalised interpretable shapelets for irregular time series |
| title_full | Generalised interpretable shapelets for irregular time series |
| title_fullStr | Generalised interpretable shapelets for irregular time series |
| title_full_unstemmed | Generalised interpretable shapelets for irregular time series |
| title_short | Generalised interpretable shapelets for irregular time series |
| title_sort | generalised interpretable shapelets for irregular time series |
| work_keys_str_mv | AT kidgerp generalisedinterpretableshapeletsforirregulartimeseries AT jamesm generalisedinterpretableshapeletsforirregulartimeseries AT lyonst generalisedinterpretableshapeletsforirregulartimeseries |