Type your query or search by category
< All Topics

What is an anomaly?

Anomaly Detection is an unsupervised Machine Learning pipeline which identifies instances in a dataset that do not conform to a regular pattern. AI Studio anomaly detector (called outlier detection in AI Studio) is an optimized implementation of the Isolation Forest Algorithm that is capable to detect anomalous instances and suspicious patterns in unlabeled datasets given a set of input fields. This means that you do not need to collect a training dataset knowing in advance which instances are anomalous and which are normal. Anomaly Detection is used for fraud detection, data cleansing Pipelines, predictive maintenance, and intrusion detection, among other examples.

Check out this video to know how to identify anomalies in data using AI Studio Outlier detection through the AI Studio Canvas.

Table of Contents