电子称重仪表决策树建模研究

针对电子称重仪表属性的多参数集对仪表性能的影响程度不同，引入了基于粗糙集理论的属性约简进行属性的降噪和排序处理，然后结合决策树理论的C4．5算法来对自诊断电子称重仪表进行分析，取信息增益率最大的结点作为决策树的根。以此使分裂信息项惩罚了多值属性最后建立了决策树模型。结果表明：此方法得到了属性的影响程度排序，使得建树快速、建模准确，利于决策分析。

Decision tree modeling of Electronic Weighing instrument

According to the noise affect from attribute set to the performance of Electronic Weighing instrument, the rough set theory based on attribute reduction noise reduction was introduced. It uses the C4.5 decision tree algorithm to modeling for self-diagnosis electronic weighing instrument and take the maximum rate of information gain as a decision tree root node, through this way, it split items of information to punish a multi-valued attribute.Finally, setting up a decision tree model. The results showed that： this approach has been the impact of sequencing properties in order to contribute modeling quicklyaccurately and facilitate decision analysis.