dotnet / dotnet/machinelearning

DetectEntireAnomalyBySrCnn - No anomalies detected as of version 1.5.4 or greater

Open
#5,891 3 comments 1 reaction 0 assignees View on GitHub
area-TimeSeries bug in-pr Priority:2
Dominant language
C#
Stars
9.4k
Forks
2k
Avg merge
2d 20h
Merged PRs (30d)
11

Description

**System Information (please complete the following information):**
- OS & Version: Windows 10
- ML.NET Version: ML.NET 1.5.4 and greater
- .NET Version: .NET 5

**Describe the bug**

When using `DetectEntireAnomalyBySrCnn` with versions 1.5.4 or greater, no anomalies are detected when running the `DetectAnomaly` method from the [phone call tutorial](https://docs.microsoft.com/en-us/dotnet/machine-learning/tutorials/phone-calls-anomaly-detection).

Changing options does nothing. Has anything changed with this algorithm that would suddenly make it stop working after 1.5.4?

If you run the sample using version 1.5.2 with the same data and options it works.

Related docs issue dotnet/docs#23888

**To Reproduce**

Follow the steps in the [phone call tutorial](https://docs.microsoft.com/en-us/dotnet/machine-learning/tutorials/phone-calls-anomaly-detection). For reference, here's a [completed code sample](https://github.com/dotnet/docs/tree/main/docs/machine-learning/tutorials/snippets/phone-calls-anomaly-detection/csharp)

**Expected behavior**

Anomalies are detected.

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.