A Fuzzy Modeling Method Based on Dirichlet Process Mixture Model for Blast Furnace Gas System

摘要

In this paper, a fuzzy modeling method based on Dirichlet process mixture model for fuzzy rules clustering is proposed. The Dirichlet process mixture model is adopted to divide the input space into several clusters, which can avoid the problem of deviating from the actual situation due to the lack of prior knowledge and operation experience. The dynamic clustering process makes the system have a certain adaptive capability during the modeling process, and the clustering judgment probability is used as the membership to determine the conclusion parameters. The experimental results using the data of blast furnace gas, which demonstrates the feasibility of the proposed model.