👍 Use an Azure AI endpoint To use an Azure AI endpoint, set thebase_url
argument:nixtla_client = NixtlaClient(base_url="you azure ai endpoint", api_key="your api_key")
threshold_method
parameter to multivariate
.
We can see that the anomalies detected for each time series occur at the
same time step since they rely on the accumulated error across all
series.
📘 In multiseries anomaly detection, error scores from all time series are aggregated at each time step, and a threshold is applied to identify significant deviations. If the aggregated error exceeds the threshold, the time step is flagged as anomalous across all series, capturing system-wide patterns.