数据挖掘方法在传统预测模型中的应用.pdfVIP

第 26 卷第 4 期 河北工业科技 Vol. 26.No. 4 2009 年 7 月 Hebei lournal of lndustrial Science and Technology luly 2009 文章编号:1008-1534(2009)04-0280-03 数据挖掘方法在传统预测模型中的应用 徐聪 (北京邮电大学软件学院，北京 102209) 摘 要:数学建模领域中，对于解决不同类型的预测问题有一些经典的预测模型。最近日益流行的 数据挖掘技术也有一些针对解决预测问题的独特的数学模型和建模方法。在此对使用传统解析式 模型的预测方法与应用数据挖掘技术的数据库模型的预测方法进行了对比，重点分析了 2 种方法 在影响因素的确定、数据规律的寻找方法以及模型侧重点选择上的不同之处，并简单讨论了如何结 合 2 种模型的优点，将数据挖掘技术应用于传统解析式模型的建摸过程之中，对已有的解析式预测 模型进行改进，进而得出更加可靠准确的预测结果。 关键词:数学建模;数据挖掘;解析式模型;数据库模型;数据统计;回归分析 中固分类号:TP311 文献标识码:A Application of data mining methods to traditional prediction mathematical models XU Cong (Software lnstitute. Beijing University of Posts and Telecommunications. Beijing 102209.China) Abstract: Many mathematical models have been developed to solve the forecasting problems. The most commonly used mathe- matical models are the traditional prediction models. Predictive mode is also one of the two basic modes in data mining meth- ods. Using the database models. thére are many particular mathematical models in data mining methods to deal with the fore- casting problem. This article made a comparison of the two kinds of prediction models in the methods of finding the influencing factors. analyzing the rule of the data and the emphasis of choosing the decision variables; and then briefly discussed how to ap- ply the data mining methods to the traditional prediction models to combine the advantages of the two to make the prediction more accurate and reliable. Key words: mathematical modeling; data mining; analytic expressions model; database model; data fitting and counting; regression analysis