Centering and scaling #688
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
vnmabus
requested changes
Jul 6, 2025
| References | ||
| ---------- | ||
|
|
||
| * J. Prothero, J. Hannig, and J. Marron. *New perspectives on centering*. The New England Journal of Statistics in Data Science, vol. 1, no. 2, 216–236, 2023. |
There was a problem hiding this comment.
We would rather use sphinxcontrib-bibtex so that all references are formatted in a consistent way.
| self.with_std = with_std | ||
| self.correction_ = correction | ||
|
|
||
| self.mean_: FData | None = None |
There was a problem hiding this comment.
The parameters that end in underscore should be set only in fit.
| msg = "Cannot center with more than one sample" | ||
| raise ValueError(msg) | ||
| result = result - center | ||
| else: |
There was a problem hiding this comment.
I do not understand the else cases. When are they applied?
| msg = "Cannot center with more than one sample" | ||
| raise ValueError(msg) | ||
| result = result - center | ||
| else: |
There was a problem hiding this comment.
I also do not understand the else cases here.
|
|
||
|
|
||
| @pytest.fixture | ||
| def sample_fdgrid() -> Generator[FDataGrid, None, None]: |
| with pytest.raises(TypeError): | ||
| root_integrated_sample_variance( | ||
| "not an FData object", # type: ignore[arg-type] | ||
| ) |
There was a problem hiding this comment.
I am missing tests for FDataBasis and for the centering and scaling methods.
Co-authored-by: Carlos Ramos Carreño <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Summary
This pull request introduces a complete framework for centering and scaling functional data in the
scikit-fdalibrary. These tools are particularly relevant in the context of vector-valued functional data and mixed datasets, where components may differ in units, scale, or variability.Motivation
In classical functional data analysis, centering and scaling are often not required, since all functions are assumed to be sampled over the same domain and measured in the same units. However, when combining multiple functional components or integrating scalar and functional features, normalization becomes critical to ensure fair comparisons and effective learning.
This PR addresses this gap by providing:
Main Additions
Transformers
CenterScaler: Flexible transformer that applies user-defined or data-driven centering and scaling operations toFDataGridorFDataBasis.StandardScaler: Computes the mean and standard deviation of the dataset and uses them to standardize functional data, mimicking scikit-learn’sStandardScaler.Summary Statistics (New Functions)
These utilities compute scalar summaries useful for centering and scaling:
individual_observation_mean: Integrated average of each function (vertical shift).grand_mean: Global scalar mean of all functions.root_integrated_sample_variance: A robust measure of variability, subtracting the mean function before integration.root_mean_square_l2: RMS of the L2 norm of each function (total magnitude without centering).individual_root_mean_square_l2: Computes RMS individually per function.These functions are available under the
skfda.exploratory.statsmodule and complement the existing set of functional location and dispersion statistics.Documentation
preprocessing/.Checklist before requesting a review