Skip to content

Decide how the kernel should be normalized. #9

Description

@abukaj

In kCSD kernel is normalised by the number of sources. That normalisation is implemented in fda73ab.

Nevertheless, multiplication of the sources' amplitude by a factor of k has exactly the same effect as multiplication of the same sources (I mean number of instances/references to them) by a factor of k^2. Should we address that somehow?

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Labels

questionFurther information is requested

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions