Add sf merlion for time-based and forecast AD #956
Conversation
oracle-contributor-agreement
bot
added
the
OCA Verified
|
|
📌 Cov diff with main: 📌 Overall coverage: |
|
|
📌 Cov diff with main: 📌 Overall coverage: |
|
codeloop
changed the title
codeloop
changed the title
|
|
codeloop
requested review from
darenr,
mayoor,
mrDzurb,
VipulMascarenhas,
qiuosier and
ahosler
as code owners
|
📌 Cov diff with main: 📌 Overall coverage: |
There was a problem hiding this comment.
-
In testing, I see that the anomaly score is no longer normalized from 0-1. Does Merlion have a different interpretation for the meaning of this score ? (CSV uploaded below)
I believe we should stick with the [0,1] convention unless we have good reason to change. -
Also Merlion seemed over-reactive on a few of the tests I've run. Are there any default settings we should be tweaking?
-
Could Merlion be made to work for non-timeseries AD? Or is it primarily time-based?
| @@ -179,7 +179,8 @@ anomaly = [ | |||
| "oracledb", | |||
| "report-creator==1.0.9", | |||
| "rrcf==0.4.4", | |||
| "scikit-learn" | |||
| "scikit-learn", | |||
| "salesforce-merlion[all]==2.0.4" | |||
There was a problem hiding this comment.
What features does the "all" get us?
| @@ -364,6 +364,7 @@ spec: | |||
| - oneclasssvm | |||
| - isolationforest | |||
| - randomcutforest | |||
| - merlion_ad | |||
There was a problem hiding this comment.
can we change this to "merlion"?
| @@ -36,7 +36,7 @@ def _build_model(self) -> AnomalyOutput: | |||
| # Set tree parameters | |||
| num_trees = model_kwargs.get("num_trees", 200) | |||
| shingle_size = model_kwargs.get("shingle_size", None) | |||
| anomaly_threshold = model_kwargs.get("anamoly_threshold", 95) | |||
| anomaly_threshold = model_kwargs.get("anomaly_threshold", 95) | |||
Add support for salesforce merlion model for time-based and forecast anomaly detection.