pruning

Terms from Artificial Intelligence: humans at the heart of algorithms

Pruning is when one removes or ignores a part of a data structure. or process. The term derives from horticultural pruning of dead or unwanted branches from a growing tree or bush. Many of AI examples are about trees, on the computational sense, hence the use in AI. One example of removal is to reduce the size of decison trees after and inductive learming algorithm such as ID3; pruning can remve branches that have low additional dscrininatiin and hence (typically) improve the generalisation of the tree. An ignoring example is during A* search where the heuristic means one knows that a particular node cannot be on the best path, and hence can prune the node from the search of the graph.

Defined on page 102

Used on pages 58, 63, 74, 79, 81, 102, 188, 355