So the branch weight is used to express the branch importance in FDT and the normalizing result of the classification weight is regarded as the confidence degree of the fault type in leave nodes.

• 用分支权表示属性分支的重要性，把分类权的正规化结果作为叶结点上各故障类型的信任度，从而提高了模糊决策树的学习精度。