honeymoose/OpenSearch
1
0
Fork 0
mirror of https://github.com/honeymoose/OpenSearch.git synced 2025-02-23 13:26:02 +00:00
Code Issues Packages Projects Releases Wiki Activity

[DOCS] Updates to outlier detection release highlight (#44911)

Browse Source
This commit is contained in:
István Zoltán Szabó 2019-07-29 08:03:11 +02:00 committed by GitHub
parent cfc8d17bb4
commit dc26521b0f
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 25 additions and 10 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
docs/reference/release-notes/highlights-7.3.0.asciidoc
View File

@ -144,16 +144,31 @@ NOTE: Data frames are only available with the default distribution of {es}.
// end::notable-highlights[]
// tag::notable-highlights[]
[float]
==== Outlier detection
[discrete]
[[release-highlights-7.3.0-outlier-detection]]
==== Discover your most unusual data using {oldetection}
{stack-ov}/security-privileges.html[Outlier detection] utilizes Elastic data
frame indexes to evaluate source indexes across multiple dimensions and identify
clusters of data based on the assigned values and which values are different
from those of the clustered data point. An outlier score can be used to indicate
how different an entity is from other entities in the index based on the
dimensions that you supply.
NOTE: Outlier detection requires a platinum license.
The goal of {stack-ov}/dfa-outlier-detection.html[{oldetection}] is to find
the most unusual data points in an index. We analyse the numerical fields of
each data point (document in an index) and annotate them with how unusual they
are.
We use unsupervised {oldetection} which means there is no need to provide a
training data set to teach {oldetection} to recognize outliers. In practice,
this is achieved by using an ensemble of distance based and density based
techniques to identify those data points which are the most different from the
bulk of the data in the index. We assign to each analysed data point an
{olscore}, which captures how different the entity is from other entities in the
index.
In addition to new {oldetection} functionality, we are introducing the
{ref}/evaluate-dfanalytics.html[evaluate {dfanalytics} API], which enables you to compute a range of performance metrics such
as confusion matrices, precision, recall, the
https://en.wikipedia.org/wiki/Receiver_operating_characteristic[receiver-operating characteristics (ROC) curve]
and the area under the ROC curve. If you are running {oldetection} on a source
index that has already been labeled to indicate which points are truly outliers
and which are normal, you can use the
evaluate {dfanalytics} API to assess the performance of the
{oldetection} analytics on your dataset.
// end::notable-highlights[]