摘要
决策树是当前预测、决策和数据挖掘中常用的方法之一。通过对决策树的生成过程进行分析,针对现有方法中决策树过度生长带来的弊端,提出了一种结合贝叶斯推理技术思想的决策树的改进方法,并给出了该方法中数据的存储结构和决策树的生成过程。该方法利用数据挖掘所产生的规则对决策树每个分支节点的分裂条件进行判断,一方面能限制决策树生长,另一方面又能帮助选择最优线路,从而使决策效率明显提高。
Decision tree is in common used on forecast, decision and data mining. In this paper, the growing process of decision trees is analyzed, and an improving method of decision tree by integrating with Bayes logic is proposed to overcome the defects from decision tree over-growth in current methods. In this method, the storage structure of data and growing technology of decision tree are given, and the split condition at each node of decision trees is judged according to rules produced in data mining. It can restrict the over-growth of decision tree and also decide the optimal route, thereby obviously promote the efficiency of decision-making.
出处
《山东建筑工程学院学报》
2003年第3期63-66,共4页
Journal of Shandong Institute of Architecture and Engineering
