Skip to content

[Algorithm] Evaluate a potential new SOTA algorithm on stream log parsing #18

@Superskyyy

Description

@Superskyyy

The following paper was recently published and claims to outperform Drain with some improvements.

Mainly the classification of logs and templates based on propositional logic.

Details 1) Proposition 1: the log message E belongs to the log template T. (2) Proposition 2: all the variable subsequences in the log message E share the same structure with their counterparts in the template T. (3) Proposition 3: not all the variable subsequences in the log message E share the same structure with their counterparts in the template T. (4) Proposition 4: the log message E does not belong to the log template T.

Although not a priority, it should be evaluated as we aim for better.
Note: this algorithm doesn't have open implementation yet. I will send an email to the authors for a replication package.
https://downloads.hindawi.com/journals/scn/2022/1559270.pdf

Metadata

Metadata

Assignees

Labels

AlgorithmThe work is on the algorithm sideResearchSome thing that highly relates to research focus, not implementation.analysis: loghelp wantedExtra attention is needed

Type

No type

Projects

Status

No status

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions