About algorithm types

The AI algorithms can be categorized into two types: ones that analyze your data in real time, and ones that must first be trained to generate an AI model.

The following are the two types of AI algorithms:

  • Trainable AI algorithms
  • Pre-trained AI algorithms

Trainable AI algorithms

These algorithms require user interaction as they must first be trained to generate an AI model. Models are well-defined computations that are formed as a result of an algorithm that takes some value, or set of values, as input and produces some value, or set of values as output. When deployed, it is your model that acts on the data and derives insights.

You must first set up training for these AI algorithms, and then train that algorithm to generate a model. The following table lists the trainable AI algorithms in alphabetical order, along with a link to the relevant training setup task.

Algorithm Action
Change risk Set up training
Metric anomaly detection Set up training
Natural language log anomaly detection Set up training
Similar tickets Set up training
Temporal grouping Set up training

Pre-trained AI algorithms

These algorithms analyze your data in real time and derive insights without the need to generate a model.

You do not need to set up training for these AI algorithms. You can enable them. In some cases, these algorithms are automatically enabled for you. The following table lists the online algorithms in alphabetical order, along with their default state, and whether you can take action on them to change that state.

Algorithm Default state Action
Probable cause Enabled None
Scope-based grouping Enabled None
Statistical baseline log anomaly detection Enabled Disable or Enable
Topological grouping Enabled None