High Quality Content by WIKIPEDIA articles! Decision tree pruning is a variant of decision tree learning, an algorithm used in data mining and machine learning. It attempts to resolve the problem of overfitting. Decision tree pruning modifies the standard learning algorithm for decision trees so that it does not split on attributes that may be irrelevant. For example, an attribute may divide the sample into two roughly equal groups. The standard decision tree learning algorithm would split on this attribute, making the tree larger but without adding predictive power. In practice, deciding whether or not to split on attribute is done using a chi-square test for statistical significance, so the tree will only split on attributes that are actually statistically significant.

Данное издание не является оригинальным. Книга печатается по технологии принт-он-деманд после получения заказа.