Skip to content

Extract the USPORF tree structure into python and use it for outlier detection #348

Description

@MaiRajborirug

Problem 1:

  • Currently, the tree structure only exists in the C++ backend and is mostly inaccessible to the user in python.

Proposal 1:

  • Create a new python/cython code that generates USPORF forest/tree structures

Problem 2:

  • There is no outlier/anomaly detection in USPORF yet

Proposal 2

  • Create an outlier detection algorithms example which contain:
  1. Isolation Forest (IF) from scikit-learn
  2. USPORF using affinity sum row of affinity matrix to determine the outliers
  3. USPORF uses pathlength function from Isolation Forest paper

Activity

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

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions