Historical Anomaly Detection Quickstart
Get started with TimeGPT’s historical anomaly detection capabilities.
Quickstart
Historical anomaly detection plot illustrating anomalies
Overview
Use the detect_anomalies
method to perform historical anomaly detection and the plot
method to visualize the identified anomalies in your time series data.
Key Concept
-
Quickly detect anomalies in your historical time series data.
-
Visualize outliers for better understanding and faster decision-making.
Steps to Detect Anomalies
1. Install and Import Packages
2. Initialize Nixtla Client
Using an Azure AI Endpoint (optional)
Using an Azure AI Endpoint (optional)
To use an Azure AI endpoint, set the base_url
argument:
3. Read the Dataset
4. Detect Anomalies
Available Models in Azure AI
Available Models in Azure AI
When using an Azure AI endpoint, set model="azureai"
:
The public API supports two models: timegpt-1
and timegpt-1-long-horizon
(default: timegpt-1
). Refer to
this tutorial to learn more about using timegpt-1-long-horizon
.
5. Plot Anomalies
6. View Logs (optional)
Processing Log Output
Processing Log Output
Next Steps
For more details about historical anomaly detection with TimeGPT, see our in-depth tutorial. This tutorial covers additional configurations and best practices for analyzing time series data at scale.