Description
Describe the feature or idea you want to propose
Tracking differentiator-based multiview dilated characteristics (TD-MVDC) is a new feature-based TSC algorithm.
The introduction of a tracking differentiator combined with dilation mapping as a preprocessor into the feature-based TSC method is proposed to improve feature diversity of TSFresh efficiently.
Ensemble feature selection based on filter feature selectors with different store ratios is designed to generate multiview features to enhance feature stability quickly.
Linear classifiers and hard voting to fastly classify and integrate multiview features to increase classification performance robustly.
The paper title is " Tracking Differentiator-based Multiview Dilated Characteristics for Time Series Classification" and has been peer-reviewed and presented at the INDIN24 conference.
The paper file and Python code can be obtained below:
https://github.com/CCHe64/TD-MVDC
Preprinted version:
https://github.com/CCHe64/TD-MVDC/blob/main/Tracking%20Differentiator-based%20Multiview%20Dilated%20Characteristics%20for%20Time%20Series%20Classification.pdf
TD-MVDC is implemented through functions in sklearn and aeon.
Python source code is here:
https://github.com/CCHe64/TD-MVDC/blob/main/Tracking%20Differentiator-based%20Multiview%20Dilated%20Characteristics.ipynb
Describe your proposed solution
Firstly, in aeon implement a new feature extraction transformation function using TSFreshEigenExtractor in transformations. collection.feature_based.
Then implement a new classification function in aeon. classification.feature_based.
Describe alternatives you've considered, if relevant
No response
Additional context
No response