Please note that the ITEA Office will be closed from 25 December 2024 to 1 January 2025 inclusive.
Patent application: Policy Induced Feature Selection
- Project
- 17002 AutoDC
- Type
- New standard
- Description
- Reduce monitoring overhead by selecting features irrespective of tasks.
- Our method does this while still being able to leverage domain expertise which helps the method in selecting generalizable features
- Contact
- Tor Björn Minde
- Tor.bjorn.minde@ericsson.com
- Technical features
Input(s):
- Measurement data.
- Policies (features that we know will be monitored).
- Number of desired features to be selected.
Main feature(s):
- Unsupervised feature selection taking domain knowledge into account.
Output(s):
- List of selected features
- Integration constraints
None, this is an intellectual property.
- Targeted customer(s)
People/software in charge of data center operations and monitoring.
- Conditions for reuse
Commercial license to be negotiated.
- Confidentiality
- Public
- Publication date
- 01-06-2020
- Involved partners
- Ericsson (SWE)
- Ericsson (Canada) (CAN)