摘要
随着数据库系统在现代企业中的广泛应用,使得企业信息化进程处于不断完善的阶段。然而,蕴藏在企业长期积累的大量数据中、对企业决策有利的“知识”并未得到有效地使用。数据挖掘技术的出现,恰恰解决了这种矛盾。目前数据挖掘技术研究应用到多个领域,涵盖的范围包括金融、电力、电信、交通等诸多领域。预测作为数据挖掘技术的一个重要组成部分也受到企业和研究学者的关注。
在当今社会中,较为准确的预测具有重要的作用,它可以帮助企业或单位作出正确的决策从而带动其发展。对于销售行业尤显重要,对药品销售企业来说正确的预测,使企业能够合理的安排企业进销存,节约成本,达到利益最大化有着更为重要的意义。尽管预测十分重要,但要得到高质量的预测结果却并非易事。如何选取合适的数据挖掘技术就成为保证预测质量的基础。
支持向量机作为数据挖掘中的新技术,基于结构风险最优化理论解决了其他技术的“过学习”现象,并运用核函数避免了“维数灾难”,并在多个领域预测中表现出不错的性能。本文主要通过对支持向量机的相关理论进行系统的阐述,针对药品销售行业,介绍了运用支持向量机的回归算法在药品销售问题上建立预测模型的步骤。使用相关软件进行仿真研究及实现预测模型,并强调了模型设计过程中销售数据的提取、预处理方法、模型验证等问题。希望对药品销售企业解决自身问题提供一定的参考。
With the wide application of database system in the modern enterprises, the enterprise informationization process is in the stage of continuous improvement. However, the "knowledge" contained in a mass of data accumulated by the enterprises, which is beneficial to enterprise decision, has not been used effectively. Data Mining is just brought to resolve this contradiction. It is currently applied to many fields covering finance, electric power, telecommunications, transportation and many others. As an important part of data mining, predication has caused the concern of the enterprises the researchers.
In modern society, a relatively accurate predication plays an important role; it helps companies and units make the right decisions and thus promote their development. It is particularly important to the marketing industry. An accurate predication is very important to Medicament sales, for it enabling enterprises to reasonable arrangements for business invoicing, saving costs, maximizing the benefits. Although the prediction is very important, but high-quality prediction is not easy. How to select appropriate data mining techniques on the basis of a prediction quality assurance.
As a new technology, Support Vector Machines in data mining based on structural risk optimization theory has solved the "over learning" phenomenon in other technology, used kernel function to avoid the "curse of dimension", and shown good predication Performance in many fields. In this paper, Support Vector Machines through a systematic exposition of the theory, for the pharmaceutical sales industry, introduces the use of Support Vector Machine regression algorithm to the problem of Medicament sales to establish prediction model of the steps. Use of relevant software simulation and implementation of forecasting models, and stressed that the model design process sales data extraction, preprocessing methods, model validation and other issues. Want to solve their own problems of Medicament sales companies to provide some reference. In this paper, through a systematic exposition of Support Vector Machines correlation theory , for the Medicament sales, introduces the steps of the construction of predication model by using Support Vector Machine regression algorithm. Used relevant software for simulation study and implementation of predication models, and stressed the methods of sales data extraction and preprocessing, model validation and other issues in model design process, Want to provide some references for solving the problems of Medicament sales.
引文
[1]张学工,统计学习理论的本质.北京:清华大学出版社,2000
[2]安金龙,支持向量机若干问题的研究.天津:天津大学,2004
[3]张学工,关于统计学习理论与支持向量机.自动化学报.2000,4
[4]李国正,支持向量机导论.北京:电子工业出版社,2004,3
[5]白鹏、张喜斌、张斌等,支持向量机理论及工程应用实例.西安:西安电子科技大学出版社,2008,8
[6]赵志英、万加桔,基于分段灰色模型的负荷预测方法.微计算机信息,2003,21
[7] Vapnik V,The Natural of Statistical Learning Theory. New York:Springer,1995.
[8] Vapnik V,Statistical Learning Theory. New York:Wiley,1998
[9] AlessandroVerri,Pattern recognition with support vector machines. first international workshop. 2002,8
[10]萧嵘、王继成、张福炎,支持向量机理论综述.计算机科学,2000,26
[11] Jiawei Han、Micheline Kamber, Data Mining Concepts and Techniques. 2001.8
[12]李昆仑,黄厚宽、田盛丰等,模糊多类支持向量机及其在入侵检测中的应用.计算机学报,2005,28
[13]饶鲜、董春曦、杨绍全,基于支持向量机的入侵检测系统.软件学报, 2003,14
[14] Suykens,Least squares support vector machines classifiers. Neural Processing,1999.3
[15] Mukkamala S、Sung AH、Abraham A,Intrusion detection using an ensemble of intelligent paradigms. Journal of Network and Computer Applications. Computational Intelligence on the Internet,2005,28
[16] Angulo.C、Parra.X,A support vector machine for multi-class classification. Neurocomputing 2003,55
[17]奉国和、朱思铭,改进及其在时间序列数据预测中的应用.华南理工大学学报:自然科学版,2005,33
[18]陈安、陈宁、周龙骧等,数据挖掘技术及应用.北京:科学出版社,2006,3
[19] Bennett K P、Bredensteiner E J,Duality and Geometry in SVM Classifiers. Proceedings of the Seventeenth International Conference on Machine Learning,2000,4
[20] Karras D A、Mertzios B G,Time series modeling of endocardial border motion in ultrasonic images comparing support vector machines,multilayer perceptrons and linear estimation techniques. Journal of the International Measurement Confederation,Imaging Measurement Systems,2004
[21]张平康、王蒙、赵登福等,基于支撑向量机的电力系统峰负荷预测.西安:西安交通大学学报,2005,39
[22]刘达、牛东晓,电力市场中电价预测模型方法及应用研究.北京:华北电力大学学报,2007,5
[23]张朝元、陈丽,一种新的支持向量回归预测模型.云南:大力学院学报,2007,6
[24]武帅、王雄等,SVM在移动通信客户流失预测中的应用研究.微计算机信息,2007,12
[25] Keerthi S S、Shevade S K、Bhattacharyya C,.A fast iterative nearest point algorithm for support vector machine classifier design. IEEE Trans Neural Netw ,2000,11
[26] Liu Zun-xiong、Zhong Hua-lan、Zhang De-yun,Multi-scale short term load prediction model using least square support vector machine. Journal of Xi’an Jiaotong University,2005,39
[27]陈春雨、林茂六、张喆，基于支持向量机的信号滤波研究.西安:西安交通大学学报,2006,40
[28] Hill S I、Wolfe P J、Rayner P J W,Nonlinear perceptual audio filtering using support vector machines. IEEE Workshop on Statistical Signal Processing Proceedings,2001
[29] Cheng Hui、Yu Qiu-ze、Tian Jin-wen,Image denoising using wavelet and support vector regression. Proceedings Third International Conference on Image and Graphics,2004
[30] Shih Pei-chung、Liu Cheng-jun,Face detection using discriminating feature analysis and support vector machine. Pattern Recog-nition,Complexity Reduction,2006,39
[31]程伟、张燕平、赵姝,支持向量机在粮食产量预测中的应用.安徽:安徽农业科学.2009,37
[32] Park S W、Park J W,Using nearest neighbor rule to improve performance of multi-class SVMs for face recognition. IEICE Transactions on Communica-tions,2004,87
[33]付岩、王耀威、王伟强等,SVM用于基于内容的自然图像分类和检索.计算机学报,2003,26
[34] Foody G M、Mathur A,A relative evaluation of multiclass image classify- cation by support vector machines. IEEE Transactions on Geoscience and Remote Sensing,2004,42
[35] Fukuda S,Katagiri R、Hirosawa H,Unsupervised approach for polarimetric SAR image classification using support vector machines. International Geoscience and Remote Sensing Symposium(IGARSS),2002,5
[36]柳新民、刘冠军、邱静,基于HMM-SVM的故障诊断模型及应用.仪器仪表学报,2006,27
[37] Bernardete R,Support vector machines for quality monitoring in a plastic injection molding process. IEEE Transactions on Systems,Man and Cyberne-tics Part C:Applications and Reviews,2005,35
[38]王国锋、刘岩、李言俊,基于支持向量机的曲线重建方法.西北工业大学学报.2004,22
[39]霍斌、白妙青,支持向量机在农业经济预测中的研究.山西:山西大学学报.2009,3
[40] An Jin-long、Wang Zheng-ou、Yang Qing-xin,et al,A SVM function approximation approach with good performances in interpolation and extrapolation. International Conference on Machine Learning and Cybernetics,2005.
[41] Chuang Chen-chia、Su,Shun-feng;Jeng Jin-tsong,Robust support vector regression networks for function approximation with outliers. IEEE Transactions on Neural Networks,2002,13
[42] de Kruif、de Vries,Comparison of four support-vector based function approximators. IEEE International Conference on Neural Networks,2004,1
[43]张莹、李凡生等,基于数据挖掘的支持向量机库存预测模型研究.中国市场.2009,9
[44]周鑫、包兴,基于LS-SVM的产品需求预测.商业现代化.2007.20
[45] Marcelino L、Ignacio S、Perez-Cruz F,Support vector regression for the simultaneous learning of a multivariate function and its derivatives. Neurocomputing,2005,69
[46]陈永义、俞小鼎、高学浩等,处理非线性分类和回归问题的一种新方法—支持向量机方法简介.应用气象学报,2004,15
[47]孙德山,支持向量机分类和回归方法研究.长沙:中南大学,2004.
[48]张恒喜、郭基联、朱家元等,小样本多元数据分析方法及应用.西安:西北工业大学出版社, 2002
[49]乐清洪、杨海瑛、朱名铨,基于有监督线性特征映射(SLFM)网络的材料性能预测模型.航空学报,2000,21
[50]陈磊、张土乔,基于最小二乘支持向量机的时用水量预测模型.哈尔滨工业大学学报, 2006, 38
[51]韩敏,混沌时间序列预测理论与方法.北京:中国水利水电出版社, 2007
[52]吕金虎、陆君安、陈士华,混沌时间序列分析及其应用.武汉:武汉大学出版社, 2002
[53]施燕杰,基于支持向量机(SVM)的股市预测方法.统计与决策, 2005, 4
[54]田翔、邓飞其,精确在线支持向量回归在股指预测中的应用.计算机工程, 2005,31
[55]周万隆、姚艳,支持向量机在股票价格短期预测中的应用.商业研究, 2006, 6
[56]王彦峰、高风,基于支持向量机的股市预测.计算机仿真, 2006, 23
[57]王裕明、吴忠,商务智能中元数据管理模型研究.计算机应用与软件, 2005, 22
[58]孙海侠,商务智能系统的构架及技术支持.情报技术, 2005
[59]元昌安,数据挖掘原理与SPSS Clementine应用.北京:电子工业出版社,2009.8
[60] DAVID H、HEIKKI M、PADHRAIC S,Principles of Data Mining [M].北京:机械工业出版社, 2003.
[61]朱幸辉、杨树强、杜凯,基于构件的海量信息存储中间件的研究与实现.计算机应用研究, 2005
[62]彭奇志,图书馆数据仓库模型的研究.现代情报,2005
[63] Efrem GMallach,决策支持与数据仓库系统[M].北京:电子工业出版社, 2001
[64]邵晨曦、潘旭伟、祁国宁等,商务智能中数据分析代理(DAA)模式的研究与实践.制造业自动化, 2004,26
[65] B.E.Boser、I.M.Guyon、and V.N.Vapnik,A Training Algorithm for Optimal MarginClassifiers. Pittsburgh,PA,1992
[66]田翔、邓飞其,精确在线支持向量回归在股指预测中的应用.计算机工程,2005,31
[67]邓乃扬、田英杰,数据挖掘中的新方法--支持向量机.北京:科学出版社,2004
[68]吴微、陈维强、刘波,用BP神经网络预测股票市场的涨跌.大连理工大学学报,2001,41