Abstract: Rough set, without providing any prior information but the necessary data sets, is a soft computing method through reducing the original data to eliminate redundant data. BP neural network is a technique which automatically forms the required decision-making region through its own learning mechanism. Considering the advantages of rough set theory and BP neural network in information disposal, this paper constructs a novel classification model of grape diseases. The results show that the model of grape diseases is feasible and effective.

Key words: rough set theory, BP neural network, grape diseases, classification model

摘要： 粗糙集无需提供问题所需处理的数据集合之外的任何先验信息，是一种通过知识约简，消除冗余数据的软计算方法；BP神经网络是一种通过自身的学习机制自动形成所要求的决策区域技术。综合了粗糙集和BP神经网络的各自优势，构建了一种新颖的葡萄病害分类模型。测试结果表明，所建模型对葡萄病害分类是行之有效的。

关键词: 粗糙集, BP神经网络, 葡萄病害, 分类模型