透气性指数回归模型的数据挖掘

Data Mining Based on Regression Model of Permeability Index

以有色冶金密闭鼓风炉的生产工艺数据为研究对象,利用SPSS(Statistical Package for the Social Science)软件,从统计分析的角度对数据进行研究;以建立密闭鼓风炉透气指数的回归模型为实例,详细描述了数据挖掘的一般流程,数据抽取、数据过滤(缺失值、异常值的处理)、数据变换(标准化)、数据建模、结果分析评价;利用统计学知识,从相关系数分析、可解释度分析、置信区间分析几个方面对建模正确性进行研究,最终从大量数据中提取出回归分析模型,以此模型,可以定量的了解变量间的相互影响,并用来预测未来因变量的变化。

Taking the technical data of imperial smelting furnace as the studying object, the paper makes statistical analysis by using SPSS. Exemplifing the permeability regression model, it fully describes the usual procedure of data analysis, including data abstraction, data filter （missing values/singular values）, data transforming （standardized）, data modeling, analysis and evaluation. By taking full use of statistical theory, the paper evaluates modeling validity through correlation coefficient analysis, variance inflation factor analysis and statistical significance analysis. Finally a regress analysis model with practical value is deducted from large amounts of above data. The model can explain the interactive influence between quantitative and variables and predict the changes of dependent variable.