Question

Which models are best suited for recursive data

Answer 1

The answer is recursive neural network.

Recursive neural network models are best suited for recursive data. They are widely used in tree structures for natural language processing and in learning sequence.

Such networks apply various constraints to represent the data. They are non-linear highly complex adaptive models that are capable of learning deep structured information. In its simplest form, two nodes are joined to form parents using a weight matrix shared across the entire network.