Problem Types

The AI & Analytics Engine platform currently supports the following three types of Machine Learning problems:

Classification

To build models to learn to predict a categorical variable from other variables, treating each row as an example or observation and columns as the variables to learn/learn from. Examples include: Default vs Non-default

Regression

To build models to learn to predict a continuous variable from other variables, treating each row as an example or observation and columns as the variables to learn/learn from.

Forecasting a Time Series

To build models that forecasts values of continuous variable(s) into the future, treating the columns as a single observation of a time-varying phenomenon characterized by numerical values.