Detection of Anomalies in High-Frequency Data

Question

How are you?

Is it right that owing to the small amount of datapoints (8640), univariate detection is not well suited for detecting abnormalities at a high frequency (millisecond sampling frequency)? Is the alternative just to add another variable and attempt multivariate detection with more variables, or is there a limit on the amount of datapoints that the inference API can handle?

We appreciate your cooperation in advance!

Answer 1

Hi, according to the documentation, "The minimum number of data points to trigger anomaly detection is 12, and the maximum is 8640 points." Furthermore, a similar question/answer has been provided here "For uni-variate detection, can you try to use length within 8640 to see how the detection performs. Actually to address the length issue, you can try sample data or aggregate data to a higher frequency. If none of the above works, would like to understand how many data points need to used for one-time detection. Multivariate detection API can not be applied to one variable. You have to prepare multiple signals to detect anomalies. For training, you could send at most 1 million points. For inference, each time you could inference 20,000 points.". Hope this helps.