Anomaly detection

[abhishekkhare19]

Power BI team recently launched a new feature of anomaly detection which automatically detects anomalies in your time series data. The only thing which we can adjust to play with data is sensitivity. You will get more anomalies data points if you increase sensitivity and vice versa. I checked entire article which explains this feature, checked the video and the algorithm(SR-CNN) behind that. I have also gone through a white paper on SR-CNN algorithm.

No where it is mentioned that how sensitivity is being used in that algorithm. How SR-CNN algorithm uses sensitivity for that purpose.

It will help us in understanding the effectiveness of process.

Any one got any idea on that? Thanks in advance.

[v-janeyg-msft · Accepted solution]

Hi, @abhishekkhare19 

The SR-CNN algorithm is an advanced and novel algorithm that is based on Spectral Residual (SR) and Convolutional Neural Network (CNN) to detect anomaly on time-series.

I think few prople know this. Roughly speaking, this function in powerbi doesn't necessarily show 100% correct results. This kind of function is the same as prediction, detecting different possible data anomalies based on sensitivity.

Anomaly detection tutorial - Power BI | Microsoft Docs

Overview of SR-CNN algorithm in Azure Anomaly Detector - Microsoft Tech Community

Best Regards

Janey Guo

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.