The mathematical details of using regression kernel for incremental learning

You state that in incremental learning using regression kernel "binary Gaussian kernel regression model for incremental learning. The kernel model maps data in a low-dimensional space into a high-dimensional space, then fits a linear model in the high-dimensional space."

I am writing a paper so I need mathematical details of mapping that maps data to high dimensional space

Answer 1

[Drew]

The documentation page that you quoted has an "Algorithms" section, and a set of references. You will likely find the answers you need in those places. See: https://www.mathworks.com/help/stats/incrementalregressionkernel.html#mw_7228730b-3b97-423b-b291-375152326425

If this answer helps you, please remember to accept the answer.

[Yasmine]

No, they don't include the details I asked about. I already checked them before asking.

[Drew]

Check the "More About" section of the corresponding doc page, fitrkernel, which has a section on "Random Feature Expansion", describing the mathematics for the feature expansion:

https://www.mathworks.com/help/stats/fitrkernel.html#mw_de8ddb55-c16b-427e-bf39-e5d3a29d2e06

For more beyond that, the references for fitrkernel and incrementalRegressionKernel have more info.

You can also use "open incrementalRegressionKernel.m" at the MATLAB prompt to read the m-file help which has a few examples referencing fitrkernel:

open incrementalRegressionKernel.m

[Yasmine]

The information I am asking about is not directly available, but I think with reading deep into the link you provided here, I will be able to extract it