Integration of CKNNRLD algorithm into Scikit-Longitudinal

[mslbiostat]

Before Submitting

• I have checked if this feature already exists or has been requested before.
• This feature aligns with the goals and scope of Scikit-Longitudinal.

What's the Problem?

Currently, Scikit-Longitudinal does not include an implementation of CKNNRLD (Cluster-based KNN Regression for Longitudinal Data), a recently proposed algorithm that enhances KNN regression performance on longitudinal datasets by incorporating a clustering-based learning mechanism.

Integrating CKNNRLD could significantly expand the library’s capabilities for regression modeling on longitudinal data, complementing existing methods primarily focused on classification tasks.

The Dream Solution

Our team would like to collaborate on integrating CKNNRLD into Scikit-Longitudinal.

We can provide:

• Algorithmic details and pseudo-code from our paper: “Boosting K-nearest neighbor regression performance for longitudinal data through a novel learning approach” (BMC Bioinformatics, 2025).
• Support in verifying correctness and maintaining consistency with the R reference implementation (currently under internal testing).

A Python version aligned with Scikit-Learn’s API could make the method more accessible and interoperable for the longitudinal data community.

Best regards,
Mohammad Sadegh Loeloe
📧 [mslbiostat@gmail.com]

Alternatives

No response

Context & Relevance

No response

Screenshots or Additional Material

No response

Willing to Contribute?

None

[simonprovost]

Hi @mslbiostat, Thanks so, so very much for opening this issue with your lab and trying to see how we can integrate CKNNRLD into Sklong :) !

First and foremost, would you prefer to give some sort of general guidance on the algorithm, where to find the R implementation precisely, etc.? Or would you prefer us to guide you on where to add a new algorithm and you to try to re-implement it in Python?

Second of all, would you confirm whether we could already in advance assume that your algorithm's authors and citation BibTex be placed in documentation once published?

Wishing you very well,

Cheers

[simonprovost]

Hi @mslbiostat

Note that I modified the doc to offer a way to properly credit everyone's algorithms and authors — as well as ongoing developments: https://scikit-longitudinal.readthedocs.io/latest/publications/

Cheers

[simonprovost]

@mslbiostat Would you be able to push here a pseudo-code of the algorithm, see if in the coming weeks I can implement it :) ?