DilationMappingTransformer
Dilation mapping transformer.
A transformer for applying an index grid dilation mapping to time series data, in the terminology of [1].
This transformation is motivated by kernel dilation, it reorders the timesteps of a time series to simulate the effect of dilation. For instance, in a pipeline, it enables a dilation-like effect for downstream models that do not inherently support such a feature.
Mathematically, the mapping operates on sequences \(x_1, \dots, x_k\). The dilation with factor \(d\) is defined as the sequence \(x_1, x_{1+d}, x_{1+2d}, \dots, x_2, x_{2+d}, x_{2+2d}, \dots, x_d, x_{2d}, \dots\), where the subsequences with grid spacing \(d\) are maximal.
The resulting sequence is of equal length to the input sequence.
This transformer reorders the values, and resets the sequence index to a RangeIndex, if the mtype is pandas based.
Quickstart
from sktime.transformations.dilation_mapping import DilationMappingTransformer
estimator = DilationMappingTransformer(dilation=2)Tags
Capabilities
- Missing values
- Unequal-length series
- Multivariate: Not supported
- Inverse transform: Not supported
- Removes missing values: Not supported
- Equalizes series length: Not supported
Properties
- Input typescitype:transform-input
- Series
- Output typescitype:transform-output
- Series
- Label typescitype:transform-labels
- None
- Fit is emptyfit_is_empty
- Yes
- Keeps the time indextransform-returns-same-time-index
- No
- Requires Xrequires_X
- Yes
- Requires yrequires_y
- No
- X and y need the same indexX-y-must-have-same-index
- No
Parameters(1)
- dilationint, default=2
- The dilation factor. Determines the spacing between original data points in the transformed series. Must be an integer greater than 0. A dilation of 1 means no change, while higher values increase the spacing.
Examples
>>> from sktime.transformations.dilation_mapping import \
... DilationMappingTransformer
>>> from sktime.datasets import load_airline
>>> y = load_airline ()
>>> y_transform = DilationMappingTransformer (dilation = 2). fit_transform (y)References
Patrick Schäfer and Ulf Leser, “WEASEL 2.0–A Random Dilated Dictionary Transform for Fast, Accurate and Memory Constrained Time Series Classification”, 2023, arXiv preprint arXiv:2301.10194.