# Machine Learning score explanation

**URL:** <https://discuss.elastic.co/t/machine-learning-score-explanation/358424>\
**Category:** Elasticsearch\
**Tags:** elastic-stack-machine-learning\
**Created:** [April 29, 2024, 11:57am UTC](https://discuss.elastic.co/t/machine-learning-score-explanation/358424 "2024-04-29T11:57:08Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Poukim0m](https://avatars.discourse-cdn.com/v4/letter/p/898d66/32.png) [@Poukim0m](https://discuss.elastic.co/u/Poukim0m)\
**Post date:** [April 29, 2024, 11:57am UTC](https://discuss.elastic.co/t/machine-learning-score-explanation/358424/1 "2024-04-29T11:57:08Z")

</div>

Hello,

according to the documentation regarding machine learning jobs the machine learning score depends on the probability that is calculated for the specific metrics and there are 3 different scores (Bucket, Influencer and records scoring) But again how is the probability calculated? And how can I drill down to events to see what caused e.g. the score decrease? I would also like to know if there is a simple way to explain what is the difference among the 3 types of scoring, cause i find it a bit confusing.

> **[Machine Learning Anomaly Scoring and Elasticsearch - How it Works](https://www.elastic.co/blog/machine-learning-anomaly-scoring-elasticsearch-how-it-works)**
>
> A complete breakdown of how machine learning in X-Pack scores anomalies and ranks them automatically on a severity scale from zero to one hundred....

Thank you in advance,  
Vivian

---

<div class="post-metadata">

**Author:** ![valeriy42](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/valeriy42/32/82758_2.png) [@valeriy42](https://discuss.elastic.co/u/valeriy42)\
**Post date:** [May 6, 2024, 3:11pm UTC](https://discuss.elastic.co/t/machine-learning-score-explanation/358424/2 "2024-05-06T15:11:54Z")

</div>

Hello @Poukim0m ,

This blog post give some more insights about what influences the record score computation. Bucket and Influence scores are the ways to aggregate record scores to provide a more general view. For instance, bucket scores would aggregate record scores from multiple anomaly detection metrics if you specified a multi-metric job.

---

<div class="post-metadata">

**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\
**Post date:** [June 3, 2024, 3:12pm UTC](https://discuss.elastic.co/t/machine-learning-score-explanation/358424/3 "2024-06-03T15:12:26Z")

</div>

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.