Self-organizing maps whose topologies can be learned with adaptive binary search trees using conditional rotations

Highlights：

• We merge the concepts of Adaptive Data Structures and Self-Organizing Maps.

• This new scheme enhances the capabilities of a Tree-based SOM, i.e., the TTOSOM.

• We attempt to preserve the topology and simultaneously find the Optimal Search Tree.

• We adapt a tree using the SOM's update rule.

• Simultaneously, we adapt a BST using rotations that are performed conditionally.