一种基于粗糙集的决策树规则提取算法

文章提出了一种新的决策树构建算法，算法选择使生成的规则满足给定可信度阀值且支持度最大的属性作为结点，不仅简化了生成的决策树；且可以根据用户输入的可信度，得到相应的决策规则集，提高了决策树的泛化能力，有效地去除了噪音规则。应用实例分析，结果表明算法提取的决策规则简洁有效。

A Rules Extraction Algorithm of Decision Tree Based on Rough Set

A new algorithm to built decision trees is proposed. The condition attribute which this algorithm choice causes the confidence degree of decision rule satisfies the value be assigned also makes the rule have biggest support degree, this method not only may simplify the production decision tree ,moreover it can get different rule set according to the confidence degree the user input ,so the extensive ability of decision tree is improved, noise rules is eliminated. The experiment shows that the algorithm provides more precise and simple decision rules.