模糊神经网络技术在纳税评估中的应用研究
|
摘要
我国税务系统在实施以综合征管、增值税专用发票稽核协查、防伪税控、出口退税、多元化电子申报纳税和行政管理办公自动化等系统为核心的信息化建设中，相继实现了省级以上的数据集中模式，初步形成了多个应用系统共享一个网络的格局。这些系统的运行产生了大量业务数据，如何将这些数据进行整合、分析和挖掘，以支持目前和今后税收的科学管理与决策，成为当前税务信息化的重要工作。因此，数据挖掘技术在税收征管领域的应用是税务部门提高执法、行政能力的必然选择。
在以往的税收纳税评估工作中，常常由于信息量不足得不到充分分析或信息过多造成冗余。针对这些问题，本文在对模糊理论和神经网络技术的基本原理及两者在数据挖掘中的优点和缺点进行了研究和分析的基础上，通过建立数学模型，将模糊规则与神经网络相结合，实现一种新的推理方法。本文通过对几种算法的分析比较，提出随机学习自动机算法在应用中具有的优势，并通过此算法进行模型设计得出仿真结果。该方法对提高基础数据的分析和利用水平，发现数据的价值具有较好的应用效果，在此基础上讨论在纳税评估应用中如何将纳税人的情况真实、准确、合法地进行审核、分析，并依法及时进行评定处理，为税务管理部门决策提供帮助。
With the rapid development in the building of information in taxation bureau, a great number of data have been generated from each respective application system, such as CTAIS, OA, etc. However, as these data are placed dispersedly, meaning differentiated, it is hard to turn these data into useful information. All of these data should be integrated, transformed, cleaned, extracted and loaded to data warehouse that help to change the data valuable. And organizations need the suitable knowledge to operate their business.
In order to serve managers more conveniently to make up their decisions and to solve problems in ratepaying evaluation in the past, this paper presents an approach, that is, making use of fuzzy neural network in data mining. This paper introduces some basic theories, models and methods of fuzzy logic, neural network, Analyses the advantage and shortage in data mining, and explains the necessary of combining the two methods. The method puts forward a new way to improve the value in using of data. By comparing and analysing several methods, this paper brings up the advantage of stochastic learning automation. According to this method, emulation result has worked out. The result shows that it has better effect in finding out data value. Based on this method, this paper discusses how to audit the information of taxpayers accurately. It helps the administration of taxation to make decisions.
在以往的税收纳税评估工作中，常常由于信息量不足得不到充分分析或信息过多造成冗余。针对这些问题，本文在对模糊理论和神经网络技术的基本原理及两者在数据挖掘中的优点和缺点进行了研究和分析的基础上，通过建立数学模型，将模糊规则与神经网络相结合，实现一种新的推理方法。本文通过对几种算法的分析比较，提出随机学习自动机算法在应用中具有的优势，并通过此算法进行模型设计得出仿真结果。该方法对提高基础数据的分析和利用水平，发现数据的价值具有较好的应用效果，在此基础上讨论在纳税评估应用中如何将纳税人的情况真实、准确、合法地进行审核、分析，并依法及时进行评定处理，为税务管理部门决策提供帮助。
With the rapid development in the building of information in taxation bureau, a great number of data have been generated from each respective application system, such as CTAIS, OA, etc. However, as these data are placed dispersedly, meaning differentiated, it is hard to turn these data into useful information. All of these data should be integrated, transformed, cleaned, extracted and loaded to data warehouse that help to change the data valuable. And organizations need the suitable knowledge to operate their business.
In order to serve managers more conveniently to make up their decisions and to solve problems in ratepaying evaluation in the past, this paper presents an approach, that is, making use of fuzzy neural network in data mining. This paper introduces some basic theories, models and methods of fuzzy logic, neural network, Analyses the advantage and shortage in data mining, and explains the necessary of combining the two methods. The method puts forward a new way to improve the value in using of data. By comparing and analysing several methods, this paper brings up the advantage of stochastic learning automation. According to this method, emulation result has worked out. The result shows that it has better effect in finding out data value. Based on this method, this paper discusses how to audit the information of taxpayers accurately. It helps the administration of taxation to make decisions.
引文
[1] McCulloch. W. S, Pitts. W. A logical calculus of the ideas immanent in Nervous Activityss [J]. biophysics, 5:115-133,1945.
[2] 李雄飞，数据挖掘与知识发现[M]，北京，高等教育出版社，2003：3-50
[3] 邵峰晶，数据挖掘原理与算法[M]，北京，中国水利水电出版社，2003，1-30，70-100
[4] 刘同明，数据挖掘技术及其应用[M]，北京，国防工业出版社，2001，1-35
[5] C. Gerior. Software Data mining[M]. DATAMATION, May 1, 1996
[6] Jiawei Han, Micheline Kamber, Data Mining: Concepts and Techniques[M], Morgan Kaufman Publishers, 2001.5-2—2.5(23):43-46.
[7] 薛荣芳、魏海鸥、耿博，应税销售额人工神经网络预测模型。税务研究，2002(11)：64-66。
[8] G. Cooper and E. Herskovits. A Bayesian method for the induction of probability networks from data. [J], UK, Machine Leaning. 1992:309-347
[9] D. Heckerman, D. Geiger, and D. Chickering. Learning Bayesian networks: The combination of knowledge and statistics data. [M], UK, Machine Learning. 1995:22-40
[10] Hebb. D.O. The Organization of Behavior [M].Wiley, 1949.
[11] Rosenblatt. F. Principles of Neurodynamics [M].Spartam, 1961.
[12] Omlin C, Giles C. Extraction of rules from discrete-time recurrent neural networks [J].Neural Networks, 9(1):41-52,1996.
[13] C. Elkan. boosting and native Bayesian learning. [J], In Technical Report CS97-557, Dept. of Computer Science and Engineering, Univ. Calif. At San Diego, Step. 1997
[14] L.A. Zadeh. Fuzzy sets[J].Information and Control, 8:338-353,1965.
[15] McCulloch. W.S, Pitts. W.A logical calculus of the ideas immanent in Nervous Activityss [J]. Biophysics, 5:115-133,1943.
[16] 孙增圻，智能控制理论与技术[M]，北京：清华大学出版社，1997；
[17] 胡宝清，模糊理论基础[M]，武汉：武汉大学出版社，2004；
[18] 武妍，施鸿宝，模糊规则自动获取及其性能的分析[J]，人工智能及应用，1025[10]：45-46，1999；
[19] Wang L.X, Mendal J.M., Fuzzy Basis Functions, Universal Appromixation, and Orthogonal Least-Souares Learnine [C]. IEEE Transcations on Neural Networks, 81.3(5):807-814,1992;
[20] Wittkemper, Hans-Georg, Steiner, Manfred. Using neural networks to forecast the systematic risk of stocks [J]. European Journal of Operational Research, 90(3):577-588,1996.
[21] Shin-ichi Horrkawa. On fuzzy modling using fuzzy neural networks with the back propagation algorithm [C]. IEEE Trans. Neural Networks, 3(5):137-155,1992;
[22] Lin chinteng etal. Neural-network-based fuzzy logic control and decision system [C]. IEEE Trans. Computers, 40(12):456-472,1991;
[23] 崔家保，税务数据仓库与数据挖掘系统研究。山东科技大学硕士论文。
[24] Mehmed Kantardzic, DATA MINING——Concepts、Models、Methods and Algorithms; Berthold, M. and D.J. Hand, Intelligent Data Analysis-An Introduction, Springer, Berlin: Germany
[25] Mamdani E. H, Assilian S. Application or fuzzy algorithms for control or simple dynamic plant[C]. IEE Proc. Part-D, 12(8), 1585-1588, 1974.
[26] T. Takagi, M. Sugeno. Fuzzy identification of systems and its applications to modelling and control [J]. IEEE Transactions on Systems, Man, and Cysbernetics, 15: 116-132, 1985
[27] 杨松林，工程模糊论方法及其应用，国防工业出版社；
[28] 王士同，模糊系统、模糊神经网络及应用程序设计，上海科学技术文献出版社。
[29] 何勇、项利国，基于模糊聚类的BP神经网络模型研究及应用，浙江大学学报；
[30] 张立明，人工神经网络模型及其应用，复旦大学出版社；
[31] [美]R_格罗思，DATA MINING—Building Competitive Advantage，
[32] 阎平凡、张长水编著，人工神经网络与模拟进化计算。清华大学出版社，2002年11月。
[33] Rumelhart. D, McClelland. J. Parallel Distributed processing, Explorations in the microstructure of cognition [M]. Cambridge: Bradford Book, MIT Press, 1-2,1986
[34] (美)George M．Marakas著，敖富江译，数据仓库、挖掘和可视化——核心概念，清华大学出版社，2004．10；
[35] (美)Pang-Ning Tan Michael Steinbach Vipin Kumar著，范明、范宏建等译，数据挖掘导论，人民邮电出版社，2006．5；
[36] Margarent H．Dunham著，郭崇慧、田凤占、靳晓明等译，陆玉昌审译，数据挖掘教程，清华大学出版社，2005．5；
[37] 毛国君、段立娟、王实、石云编著，数据挖掘管理与算法，清华大学出版社，2005．7；
[38] 梁循，数据挖掘算法与应用，北京大学出版社，2006．4；
[39] (加)Jiawer Han Micheline Kamber著，范明、孟小峰译，数据挖掘挖掘与技术，机械工业出版社，2006．3；
[40] 王小平，曹立明．遗传算法理论、设计及软件实现[M]．西安：西安交通大学出版社，2002．
[41]] 金人庆，中国当代税收要论。人民出版社。2002年6月。
[42] 王卫红、陈健，税务稽查选案决策支持系统的开发。计算机工程与应用，2000(5)：186-189。
[43] 王锐，税收不遵从的识别研究。浙江大学博士论文。
[44] Dorian Pyle著，杨冬青、马秀莉、唐世谓译，业务建模与数据挖掘，2005．4；
[45] 郑未名，模糊神经网络的学习算法。武汉大学硕士论文。
[46] 王振雷，模糊神经网络理论及其在复杂系统中的应用研究。东北大学博士论文。
[47] 张伦俊．税收与GDP等主要经济指标的数量关系．税务研究，2005，7：11～14。
[2] 李雄飞，数据挖掘与知识发现[M]，北京，高等教育出版社，2003：3-50
[3] 邵峰晶，数据挖掘原理与算法[M]，北京，中国水利水电出版社，2003，1-30，70-100
[4] 刘同明，数据挖掘技术及其应用[M]，北京，国防工业出版社，2001，1-35
[5] C. Gerior. Software Data mining[M]. DATAMATION, May 1, 1996
[6] Jiawei Han, Micheline Kamber, Data Mining: Concepts and Techniques[M], Morgan Kaufman Publishers, 2001.5-2—2.5(23):43-46.
[7] 薛荣芳、魏海鸥、耿博，应税销售额人工神经网络预测模型。税务研究，2002(11)：64-66。
[8] G. Cooper and E. Herskovits. A Bayesian method for the induction of probability networks from data. [J], UK, Machine Leaning. 1992:309-347
[9] D. Heckerman, D. Geiger, and D. Chickering. Learning Bayesian networks: The combination of knowledge and statistics data. [M], UK, Machine Learning. 1995:22-40
[10] Hebb. D.O. The Organization of Behavior [M].Wiley, 1949.
[11] Rosenblatt. F. Principles of Neurodynamics [M].Spartam, 1961.
[12] Omlin C, Giles C. Extraction of rules from discrete-time recurrent neural networks [J].Neural Networks, 9(1):41-52,1996.
[13] C. Elkan. boosting and native Bayesian learning. [J], In Technical Report CS97-557, Dept. of Computer Science and Engineering, Univ. Calif. At San Diego, Step. 1997
[14] L.A. Zadeh. Fuzzy sets[J].Information and Control, 8:338-353,1965.
[15] McCulloch. W.S, Pitts. W.A logical calculus of the ideas immanent in Nervous Activityss [J]. Biophysics, 5:115-133,1943.
[16] 孙增圻，智能控制理论与技术[M]，北京：清华大学出版社，1997；
[17] 胡宝清，模糊理论基础[M]，武汉：武汉大学出版社，2004；
[18] 武妍，施鸿宝，模糊规则自动获取及其性能的分析[J]，人工智能及应用，1025[10]：45-46，1999；
[19] Wang L.X, Mendal J.M., Fuzzy Basis Functions, Universal Appromixation, and Orthogonal Least-Souares Learnine [C]. IEEE Transcations on Neural Networks, 81.3(5):807-814,1992;
[20] Wittkemper, Hans-Georg, Steiner, Manfred. Using neural networks to forecast the systematic risk of stocks [J]. European Journal of Operational Research, 90(3):577-588,1996.
[21] Shin-ichi Horrkawa. On fuzzy modling using fuzzy neural networks with the back propagation algorithm [C]. IEEE Trans. Neural Networks, 3(5):137-155,1992;
[22] Lin chinteng etal. Neural-network-based fuzzy logic control and decision system [C]. IEEE Trans. Computers, 40(12):456-472,1991;
[23] 崔家保，税务数据仓库与数据挖掘系统研究。山东科技大学硕士论文。
[24] Mehmed Kantardzic, DATA MINING——Concepts、Models、Methods and Algorithms; Berthold, M. and D.J. Hand, Intelligent Data Analysis-An Introduction, Springer, Berlin: Germany
[25] Mamdani E. H, Assilian S. Application or fuzzy algorithms for control or simple dynamic plant[C]. IEE Proc. Part-D, 12(8), 1585-1588, 1974.
[26] T. Takagi, M. Sugeno. Fuzzy identification of systems and its applications to modelling and control [J]. IEEE Transactions on Systems, Man, and Cysbernetics, 15: 116-132, 1985
[27] 杨松林，工程模糊论方法及其应用，国防工业出版社；
[28] 王士同，模糊系统、模糊神经网络及应用程序设计，上海科学技术文献出版社。
[29] 何勇、项利国，基于模糊聚类的BP神经网络模型研究及应用，浙江大学学报；
[30] 张立明，人工神经网络模型及其应用，复旦大学出版社；
[31] [美]R_格罗思，DATA MINING—Building Competitive Advantage，
[32] 阎平凡、张长水编著，人工神经网络与模拟进化计算。清华大学出版社，2002年11月。
[33] Rumelhart. D, McClelland. J. Parallel Distributed processing, Explorations in the microstructure of cognition [M]. Cambridge: Bradford Book, MIT Press, 1-2,1986
[34] (美)George M．Marakas著，敖富江译，数据仓库、挖掘和可视化——核心概念，清华大学出版社，2004．10；
[35] (美)Pang-Ning Tan Michael Steinbach Vipin Kumar著，范明、范宏建等译，数据挖掘导论，人民邮电出版社，2006．5；
[36] Margarent H．Dunham著，郭崇慧、田凤占、靳晓明等译，陆玉昌审译，数据挖掘教程，清华大学出版社，2005．5；
[37] 毛国君、段立娟、王实、石云编著，数据挖掘管理与算法，清华大学出版社，2005．7；
[38] 梁循，数据挖掘算法与应用，北京大学出版社，2006．4；
[39] (加)Jiawer Han Micheline Kamber著，范明、孟小峰译，数据挖掘挖掘与技术，机械工业出版社，2006．3；
[40] 王小平，曹立明．遗传算法理论、设计及软件实现[M]．西安：西安交通大学出版社，2002．
[41]] 金人庆，中国当代税收要论。人民出版社。2002年6月。
[42] 王卫红、陈健，税务稽查选案决策支持系统的开发。计算机工程与应用，2000(5)：186-189。
[43] 王锐，税收不遵从的识别研究。浙江大学博士论文。
[44] Dorian Pyle著，杨冬青、马秀莉、唐世谓译，业务建模与数据挖掘，2005．4；
[45] 郑未名，模糊神经网络的学习算法。武汉大学硕士论文。
[46] 王振雷，模糊神经网络理论及其在复杂系统中的应用研究。东北大学博士论文。
[47] 张伦俊．税收与GDP等主要经济指标的数量关系．税务研究，2005，7：11～14。