自动生成决策树的通用算法模板

目的 提供自动生成决策树的通用算法模板，从而使算法的设计能够快速验证为解决特定决策问题而设计的新算法。方法 提供智能编辑器供算法设计定义自己的公式，并利用该公式或内置算示来初始化通用算法模板，从而实现连续属性的离散化、采用基于概率分布的方法处理未知属性和噪声数据，利用预剪枝和后剪枝算法进行剪枝。结果与结论 用户能够利用该模板，在系统提供的交互式图形环境中，针对不同的决策问题测试决策树生成算法。

A General Algorithm Template for Automatic Generation of Decision Trees

Aim To provide a general algorithm template for automatic generation of decision trees, making the designers of the decision tree algorithms be able to use this template to evaluate the new algorithms for solving specific decision problems. Methods The general template is initialized by the algorithm designers' own fomula edited in the intelligent editor provided by this paper or the built in algorithms, thus the continuous attributes are discretized. The unknown attributes and the noise data are processed with probability, the pre pruning and post pruning are used to prune the decision tree. Results and Conclusion Using this template, the new algorithms can be tested on different decision problems in the interactive graphic environment provided by this system.