Anomaly detection

Anomaly detection (or outlier analysis) is a step in data mining that identifies data points, events, and / or observations that deviate from the normal behavior of a data set.

With all the analytics and management software available, businesses can now more easily than ever measure every aspect of their business operations effectively. This includes the operational performance of applications and infrastructure components, as well as key performance indicators (KPIs) that measure the success of an organization. With millions of metrics that can be measured, companies usually have a fairly impressive set of data to study the performance of their business.

This dataset contains data patterns that represent normal business activities. An unexpected change within these data patterns or an event that does not match the expected data pattern is considered an anomaly. In other words, an anomaly is a deviation from normal practice.

Anomalous data may indicate critical incidents such as a technical failure or potential opportunities such as a change in consumer behavior.

No change can be an anomaly if it breaks a pattern that is normal for data from that particular metric. Anomalies are not categorically good or bad, they are just deviations from the expected value of a metric at any given time

Machine learning is used to automate anomaly detection, which improves the quality of detection and reduces the number of unnecessary alerts.

Examples of anomaly detection applications:

Our solution allows for detection and definition

Setting key points
The user enters his own set of parameters, after exceeding any of them, the system will send a notification
Detecting anomalies in the data
The user enters data samples of the data, and the neural network learns to automatically detect data that does not meet the criteria
Data over time
After determining the expected amount of data in time (e.g. records per hour), the system monitors whether the entered data complies with the specified criterion.
Own conditions
After determining the expected amount of data in time (e.g. records per hour), the system monitors whether the entered data complies with the specified criterion.



<