一个增量式判定树学习算法INDUCE

INDUCE: AN INCREMENTAL DECISION TREE ALGORITHM

ＩＮＤＵＣＥ算法采用自顶向下判定树归纳的学习方法，不仅具有健壮性好、效率高和正确率高等优点，还具有增量学习能力，可以动态修正概念描述的不足．该算法还运用了构造性归纳的思想，在学习过程中生成新的描述子，使概念描述空间搜索的效率得到提高．运行实例表明，ＩＮＤＵＣＥ具有很好的应用前景．

INDUCE is a version of top\|down induction of decision trees. It has the merit of being robust, efficient and precise, and it is capable of incremental learning, so as to modify concept description dynamically. INDUCE also employs the idea of constructive induction, which improves the efficiency of searching in concept description space by generating new descriptors. Successful applications of INDUCE indicate that it has a bright future.