Advanced Analytics applications often require data models based on analytical base tables: tables that include either measures and dimensional attributes at the same time. This kind of tables contain all the quantitative and categorical features needed by machine learning algorithms to fit predictive and analytical models.

Analytical base tables are fact tables that embed the columns of the related dimensional tables instead of referencing them.

The new BI Modeler release (b115) implements a new Hierarchy Implementation Strategy to generate, starting from a standard Fact Schema, this kind of fact tables with embedded dimensional attributes.

Applying this implementation strategy, called Embedded Dimension, a Hierarchy of a Fact Schema will be implemented adding all the columns derived from its attributes to the corresponding fact table, side by side with the measures of the Fact Schema. The grain of the fact table is always defined by the set of dimensions of every Hierarchy.