Based On Data Mining Technology Of The Imperial Smelting Furnace Ventilation Situation Analysis

With the quick development of database technique and abroad use of DBMS, people have accumulated more and more data. The richness brings the demand for powerful data analysis tool. Under these circumstances, Data Mining (DM) technique appears and gradually obtains the wide application in business, construction, scientific research and so on.Imperial Smelting Furnace (ISF) is the nonlinear,time-various and strongly coupling system, whose operation mechanism is complex and mode is difficult to be established. Permeability is one of important index for ISF production, and the accurate analysis for it is significant to increase the output. ISF Field collection need numerous checking posts and will produce huge data, therefore, the research about how to effectively use data analysis tool to find out the potent and valuable information is highly valuable and can direct the production practices.The dissertation focuses on the permeability in the smelting of ISF. Based on statistical analysis and "decision tree" theory, it researches the permeability by using the Data Mining technology.Firstly, the dissertation statistically examines the production data of the Zinc-Lead smelting process of IBF. Through data pretreatment,compression and relative analyses, it determines four primary elements of effecting permeability and produces multivariate linear regression model. Considering the nonlinearity of ISF production, the dissertation introduces the algorithm of self-adaptable fading weight and makes emulation tests for the effectivity of permeability analysis mode to make the linear model to be better forecast to the permeability.Then, the dissertation explores the production data by using algorithm based on "decision tree" theory. Through "information gain" defined by information entropy, it chooses nodes, produces "decision tree" and abstracts analysis rules. Since the rule sets are very huge, the dissertation simplifies the "decision tree" rules by using PEP pruning algorithm. By using database technique, the dissertation establishes the "decision tree" for one-year production data and examines the permeability analysis mode. After using the PEP pruning algorithm, the "decision tree" is provided with excellent analysis effects.