Skip to content

UKF log-likelihood computation is likely wrong #68

Description

@GeorgePearse

Original issue: rlabbe/filterpy#114
Created: 2018-04-29
Last updated: 2018-04-29


I just ran the example in the UKF.py docstring, and the log-likelihood are in the range of -20000 or so, Either the computation is incorrect or the problem formulation is wrong.

Metadata

Metadata

Assignees

No one assigned

    Labels

    from-upstreamIssue imported from upstream rlabbe/filterpy repository

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions