A method to model dependencies between data streams sequentially in blocks, reducing computational complexity while capturing inter-stream relationships.