Research on fault diagnosis method of blast furnace based on clustering combine SVMs dynamic pruned binary tree

Since fault diagnosis of blast furnace is very important in manufacturing, in this paper, a new strategy based on clustering combining SVMs pruned binary tree is proposed to solve diagnosis problem in blast furnace. According to the relations of categories in multi-class problem, it is needless to distinguish all the sorts. In order to improve classification efficiency, advantage of clustering and support vector machine is combined. According to the similarity of different samples´ sorts, a binary tree is constructed rationally to accelerate fault diagnosis efficiency. The class similarity is determine according to class distance and distribution sphere in feature space, the similarity is used to determine the classification order of hierarchical multi-class classify SVMs. The training samples and corresponding SVMs sub-classifiers are selectively re-constructed to make sure bigger classification margin and good generalization ability. The results of simulation experiments show that the proposed method is faster in training and classifying, better in classification correctness and generalization.