• tor l of products in a factory [3]. Incremental processes analyze data using preliminary views and then, narrow down the views to important candidate issues in order to identify in- sights in data. For example, important characteristics of analyses requires techniques associated with analysis of planning and management. In these contexts, data mining techniques together with their applications to biology, chemistry and ecology [6,7]. Some other papers apply data mining to evaluation tasks needed in product management and material recycling [8,9]. Applications of evaluation and requirement patterns for marketing [14]. Methodologies employed are case-based reasoning, decision rule learning atic ADVANCED ENGINEERING INFORMATICS diseases in evidence-based medicine have been identified in this way [4]. Iterations of data analysis under diversified views are used to support idea creation to develop new products and services [5]. The design of such meta-level management in system identification are also seen [10,11]. Some recent work applied data mining to design support in architecture and manufacturing fields [12,13]. The journal also includes a paper on acquiring customer Edi Applications eligib With progress in information and communication tech- nology, voluminous records of scientific, engineering and social activities have been accumulated in computer net- works. Data mining techniques attract much industrial attention since they enable development of powerful tools to extract knowledge from data [1,2]. Information retrieval and statistical analysis are the conventional techniques relevant to data mining, where the former is to seek specific knowledge and facts that are targeted by users, and the latter is to identify trends in data using probabilistic theory and statistics. In general, information retrieval relies on efficient search, whereas statistical analysis processes limited amounts of data, and uses sampling techniques to reduce computational costs if necessary. In contrast, mod- ern data mining applications involve different analysis objectives. The main objective is often to discover knowl- edge of trends that are embedded in data. However, mod- ern data mining applications do not exclude information retrieval and statistical analyses. Data mining often uses these techniques in combination with machine learning methodologies to analyze data. Data mining processes consist of tasks such as cleaning/ integration/selection/transformation of data, and presenta- tion/evaluation of analysis results [1]. These are interac- tively repeated until analysis goals are achieved. The success of data mining depends on the meta-level design of users’ work processes. Top–down and hierarchical decomposition of users’ analysis goals and applications of data mining along this hierarchy are used under clear and concrete objectives such as those governing quality control Advanced Engineering Inform 1474-0346/$ - see front matter � 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.aei.2007.01.001 ial e for data mining involves synthetic engineering to employ a range of tech- niques appropriately in order to achieve objectives. Recently, we have observed successful applications of data mining that are based on well-designed data mining processes in various industries. For example, candidate customers who have high sensitivity to direct mail advertise- ments are efficiently extracted from data of customers’ behavior in the field of marketing. Knowledge extraction to find loyal customers for sales promotions is attempted in applications based on point-of-sales (POS) data. The characterization of web-site visitors is another popular application of data mining in marketing. In finance, credit risks of individuals and companies are successfully evalu- ated based on their attribute records. The detection of can- didate fraud transactions in credit card records is another successful application. A major application of data mining in manufacturing is quality control. For example, monitor- ing the production yield of semi-conductor LSI and diagno- sis of its degradation is an application that has resulted in cost reductions. Defect detection and diagnosis of consumer products is another application that is crucial to product liability. However, compared with service industries, data mining has been used in limited fields in manufacturing. This is because the grasp of entire trends governing product manufacture is a major task compared with the capture of segmented characteristics in other applications. The volumes of Advanced Engineering Informatics (AEI) in the last five years include approximately ten papers related to techniques and applications of data min- ing. Papers propose novel and basic data mining www.elsevier.com/locate/aei s 21 (2007) 241–242
  • and artificial neural networks. All of these techniques are well-known and are supported by widely applicable tools due to their long research history. Particularly, case-based reasoning consequences and responsible use. Technology (EPFL), Switzerland and T. Tomiyama, Intel- ligent Mechanical Systems, Delft University of Techno- logy, Netherlands for their extensive supports to complete this Special Issue. 242 Editorial / Advanced Engineering Informatics 21 (2007) 241–242 Papers describing promising applications of novel data mining techniques have been presented in many interna- tional conferences and workshops on the data mining. Conferences and journals dedicated to each application domain also contain significant numbers of data mining applications. However, domain crossing publications focusing on technical achievements in applications rather than technical novelty are not available. Such publications will clarify not only useful data mining techniques for various domains but also techniques are identified for future study of data mining. This Special Issue on applica- tions eligible for data mining will provide such clarification. Also, this issue is intended to promote development of data mining applications in general. Five papers are included in this Special Issue. The first paper, authored by Jan Ramon et al., presents the applica- tion of data mining methods to predict the evolution of patients in an intensive care unit. The second paper of Yufei Shu presents the application to a nonlinear structural health inference technique for nuclear power plants based on interactive data mining. The third paper, authored by Yu-Chiang Lia et al., addresses the privacy preserving problem of finding optimal sanitization methods where all restrictive itemsets are concealed while minimizing costs. The fourth paper by Arijit Laha discusses a data driven method of building credit scoring models. The final paper authored by Sung Ho Ha presents a marketing analysis framework based on customer segmentation and a cus- tomer model of the transition among segments. The data mining techniques that are used in these papers share char- acteristics with research that has been reported in this journal over the last five years. Many of them employ rule learning and neural network frameworks. The interpret- ability of the outcome, the applicability and the efficient implementation are important issues for data mining appli- cations, and these issues related to synthetic engineering should be addressed in future studies of data mining. Acknowledgments I express my gratitude to I.F.C. Smith in Applied Com- puting and Mechanics (IMAC), Swiss Federal Institute of reasoning and decision rule learning provide interpretable grounds for reasoning outcomes in form of past cases and rules. These technical characteristics are keys for suc- cessful applications of data mining since they are advanta- geous for efficient implementation, validity checking of References [1] Usama Fayyad, From data mining to knowledge discovery in database, AI Magazine 17 (3) (1996) 37–54. [2] Jiawei Han, Micheline Kamber, Data Mining, Concepts and Tech- niques, second ed. The Morgan Kaufmann Series in Data Management Systems, Morgan Kaufmann Publishers, 2006. [3] Satoshi Hori, A Watchdog System for Field Quality – A Basket Analysis Approach, in: Proceedings of US–Japan FA Symposium, vol. 2, 1998, pp. 741–748. [4] Hhiroshi Motoda et al., Application of DT-GBI to Promoter and Hepatitis Datasets, in: Proceedings of the Knowledge Discovery in BioMedicine, KDbM-04, 2004, pp. 10–40. [5] Yukio Ohsawa, Chance discoveries for making decisions in complex real world, New Generation Computing 20 (2) (2002) 143–163. [6] Takashi Matsuda, Hiroshi Motoda and Takashi Washio: Graph-based induction and its applications, Advanced Engineering Informatics 16 (2) (2002) 135–143. [7] Colin Fyfe, Juan Corchado, A comparison of Kernel methods for instantiating case based reasoning systems, Advanced Engineering Informatics 16 (3) (2002) 165–178. [8] Darren Graham, Simon D. Smith, Estimating the productivity of cyclic construction operations using case-based reasoning, Advanced Engi- neering Informatics 18 (1) (2004) 17–28. [9] Li-Hsing Shih, Yu-Si Chang, Yung-Teh Lin, Intelligent evaluation approach for electronic product recycling via case-based reasoning, Advanced Engineering Informatics 20 (2) (2006) 137–145. [10] Hai Qiu, Jay Lee, Jing Lin, Gang Yu, Robust performance degradation assessment methods for enhanced rolling element bearing prognostics, Advanced Engineering Informatics 17 (3-4) (2003) 127– 140. [11] S. Saitta, B. Raphael, I.F.C. Smith, Data mining techniques for improving the reliability of system identification, Advanced Engi- neering Informatics 19 (4) (2005) 289–298. [12] Baranidharan Raman, Jody R. Naderi, Computer based pedestrian landscape design using decision tree templates, Advanced Engineering Informatics 20 (1) (2006) 23–30. [13] Yoko Ishino, Yan Jin, An information value based approach to design procedure capture, Advanced Engineering Informatics 20 (1) (2006) 89–107. [14] Chun-Hsien Chen, Li Pheng Khoo, Wei Yan, A strategy for acquiring customer requirement patterns using laddering technique and ART2 neural network, Advanced Engineering Informatics 16 (3) (2002) 229– 240. Takashi Washio The Institute of Scientific and Industrial Research (ISIR), Osaka University, 8-1, Mihogaoka, Ibaraki City, Osaka 567-0047, Japan E-mail address: washio@ar:sanken:osaka-u:ac:jp Applications eligible for data mining Acknowledgments References
Please download to view
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
...

Applications eligible for data mining

by takashi-washio

on

Report

Category:

Documents

Download: 0

Comment: 0

212

views

Comments

Description

Download Applications eligible for data mining

Transcript

  • tor l of products in a factory [3]. Incremental processes analyze data using preliminary views and then, narrow down the views to important candidate issues in order to identify in- sights in data. For example, important characteristics of analyses requires techniques associated with analysis of planning and management. In these contexts, data mining techniques together with their applications to biology, chemistry and ecology [6,7]. Some other papers apply data mining to evaluation tasks needed in product management and material recycling [8,9]. Applications of evaluation and requirement patterns for marketing [14]. Methodologies employed are case-based reasoning, decision rule learning atic ADVANCED ENGINEERING INFORMATICS diseases in evidence-based medicine have been identified in this way [4]. Iterations of data analysis under diversified views are used to support idea creation to develop new products and services [5]. The design of such meta-level management in system identification are also seen [10,11]. Some recent work applied data mining to design support in architecture and manufacturing fields [12,13]. The journal also includes a paper on acquiring customer Edi Applications eligib With progress in information and communication tech- nology, voluminous records of scientific, engineering and social activities have been accumulated in computer net- works. Data mining techniques attract much industrial attention since they enable development of powerful tools to extract knowledge from data [1,2]. Information retrieval and statistical analysis are the conventional techniques relevant to data mining, where the former is to seek specific knowledge and facts that are targeted by users, and the latter is to identify trends in data using probabilistic theory and statistics. In general, information retrieval relies on efficient search, whereas statistical analysis processes limited amounts of data, and uses sampling techniques to reduce computational costs if necessary. In contrast, mod- ern data mining applications involve different analysis objectives. The main objective is often to discover knowl- edge of trends that are embedded in data. However, mod- ern data mining applications do not exclude information retrieval and statistical analyses. Data mining often uses these techniques in combination with machine learning methodologies to analyze data. Data mining processes consist of tasks such as cleaning/ integration/selection/transformation of data, and presenta- tion/evaluation of analysis results [1]. These are interac- tively repeated until analysis goals are achieved. The success of data mining depends on the meta-level design of users’ work processes. Top–down and hierarchical decomposition of users’ analysis goals and applications of data mining along this hierarchy are used under clear and concrete objectives such as those governing quality control Advanced Engineering Inform 1474-0346/$ - see front matter � 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.aei.2007.01.001 ial e for data mining involves synthetic engineering to employ a range of tech- niques appropriately in order to achieve objectives. Recently, we have observed successful applications of data mining that are based on well-designed data mining processes in various industries. For example, candidate customers who have high sensitivity to direct mail advertise- ments are efficiently extracted from data of customers’ behavior in the field of marketing. Knowledge extraction to find loyal customers for sales promotions is attempted in applications based on point-of-sales (POS) data. The characterization of web-site visitors is another popular application of data mining in marketing. In finance, credit risks of individuals and companies are successfully evalu- ated based on their attribute records. The detection of can- didate fraud transactions in credit card records is another successful application. A major application of data mining in manufacturing is quality control. For example, monitor- ing the production yield of semi-conductor LSI and diagno- sis of its degradation is an application that has resulted in cost reductions. Defect detection and diagnosis of consumer products is another application that is crucial to product liability. However, compared with service industries, data mining has been used in limited fields in manufacturing. This is because the grasp of entire trends governing product manufacture is a major task compared with the capture of segmented characteristics in other applications. The volumes of Advanced Engineering Informatics (AEI) in the last five years include approximately ten papers related to techniques and applications of data min- ing. Papers propose novel and basic data mining www.elsevier.com/locate/aei s 21 (2007) 241–242
  • and artificial neural networks. All of these techniques are well-known and are supported by widely applicable tools due to their long research history. Particularly, case-based reasoning consequences and responsible use. Technology (EPFL), Switzerland and T. Tomiyama, Intel- ligent Mechanical Systems, Delft University of Techno- logy, Netherlands for their extensive supports to complete this Special Issue. 242 Editorial / Advanced Engineering Informatics 21 (2007) 241–242 Papers describing promising applications of novel data mining techniques have been presented in many interna- tional conferences and workshops on the data mining. Conferences and journals dedicated to each application domain also contain significant numbers of data mining applications. However, domain crossing publications focusing on technical achievements in applications rather than technical novelty are not available. Such publications will clarify not only useful data mining techniques for various domains but also techniques are identified for future study of data mining. This Special Issue on applica- tions eligible for data mining will provide such clarification. Also, this issue is intended to promote development of data mining applications in general. Five papers are included in this Special Issue. The first paper, authored by Jan Ramon et al., presents the applica- tion of data mining methods to predict the evolution of patients in an intensive care unit. The second paper of Yufei Shu presents the application to a nonlinear structural health inference technique for nuclear power plants based on interactive data mining. The third paper, authored by Yu-Chiang Lia et al., addresses the privacy preserving problem of finding optimal sanitization methods where all restrictive itemsets are concealed while minimizing costs. The fourth paper by Arijit Laha discusses a data driven method of building credit scoring models. The final paper authored by Sung Ho Ha presents a marketing analysis framework based on customer segmentation and a cus- tomer model of the transition among segments. The data mining techniques that are used in these papers share char- acteristics with research that has been reported in this journal over the last five years. Many of them employ rule learning and neural network frameworks. The interpret- ability of the outcome, the applicability and the efficient implementation are important issues for data mining appli- cations, and these issues related to synthetic engineering should be addressed in future studies of data mining. Acknowledgments I express my gratitude to I.F.C. Smith in Applied Com- puting and Mechanics (IMAC), Swiss Federal Institute of reasoning and decision rule learning provide interpretable grounds for reasoning outcomes in form of past cases and rules. These technical characteristics are keys for suc- cessful applications of data mining since they are advanta- geous for efficient implementation, validity checking of References [1] Usama Fayyad, From data mining to knowledge discovery in database, AI Magazine 17 (3) (1996) 37–54. [2] Jiawei Han, Micheline Kamber, Data Mining, Concepts and Tech- niques, second ed. The Morgan Kaufmann Series in Data Management Systems, Morgan Kaufmann Publishers, 2006. [3] Satoshi Hori, A Watchdog System for Field Quality – A Basket Analysis Approach, in: Proceedings of US–Japan FA Symposium, vol. 2, 1998, pp. 741–748. [4] Hhiroshi Motoda et al., Application of DT-GBI to Promoter and Hepatitis Datasets, in: Proceedings of the Knowledge Discovery in BioMedicine, KDbM-04, 2004, pp. 10–40. [5] Yukio Ohsawa, Chance discoveries for making decisions in complex real world, New Generation Computing 20 (2) (2002) 143–163. [6] Takashi Matsuda, Hiroshi Motoda and Takashi Washio: Graph-based induction and its applications, Advanced Engineering Informatics 16 (2) (2002) 135–143. [7] Colin Fyfe, Juan Corchado, A comparison of Kernel methods for instantiating case based reasoning systems, Advanced Engineering Informatics 16 (3) (2002) 165–178. [8] Darren Graham, Simon D. Smith, Estimating the productivity of cyclic construction operations using case-based reasoning, Advanced Engi- neering Informatics 18 (1) (2004) 17–28. [9] Li-Hsing Shih, Yu-Si Chang, Yung-Teh Lin, Intelligent evaluation approach for electronic product recycling via case-based reasoning, Advanced Engineering Informatics 20 (2) (2006) 137–145. [10] Hai Qiu, Jay Lee, Jing Lin, Gang Yu, Robust performance degradation assessment methods for enhanced rolling element bearing prognostics, Advanced Engineering Informatics 17 (3-4) (2003) 127– 140. [11] S. Saitta, B. Raphael, I.F.C. Smith, Data mining techniques for improving the reliability of system identification, Advanced Engi- neering Informatics 19 (4) (2005) 289–298. [12] Baranidharan Raman, Jody R. Naderi, Computer based pedestrian landscape design using decision tree templates, Advanced Engineering Informatics 20 (1) (2006) 23–30. [13] Yoko Ishino, Yan Jin, An information value based approach to design procedure capture, Advanced Engineering Informatics 20 (1) (2006) 89–107. [14] Chun-Hsien Chen, Li Pheng Khoo, Wei Yan, A strategy for acquiring customer requirement patterns using laddering technique and ART2 neural network, Advanced Engineering Informatics 16 (3) (2002) 229– 240. Takashi Washio The Institute of Scientific and Industrial Research (ISIR), Osaka University, 8-1, Mihogaoka, Ibaraki City, Osaka 567-0047, Japan E-mail address: washio@ar:sanken:osaka-u:ac:jp Applications eligible for data mining Acknowledgments References
Fly UP