Applications Of Data Mining In Ebusiness And Finance, Frontiers in Artificial Intelligence and Applications, Volume 177

The application of Data Mining (DM) technologies has shown an explosive growth in an increasing number of different areas of business, government and science. Two of the most important business areas are finance, in particular in banks and insurance companies, and e-business, such as web portals, e-commerce and ad management services.In spite of the close relationship between research and practice in Data Mining, it is not easy to find information on some of the most important issues involved in real world application of DM technology, from business and data understanding to evaluation and deployment. Papers often describe research that was developed without taking into account constraints imposed by the motivating application. When these issues are taken into account, they are frequently not discussed in detail because the paper must focus on the method. Therefore knowledge that could be useful for those who would like to apply the same approach on a related problem is not shared. The papers in this book address some of these issues. This book is of interest not only to Data Mining researchers and practitioners, but also to students who wish to have an idea of the practical issues involved in Data Mining.IOS Press is an international science, technical and medical publisher of high-quality books for academics, scientists, and professionals in all fields. Some of the areas we publish in: -Biomedicine -Oncology -Artificial intelligence -Databases and information systems -Maritime engineering -Nanotechnology -Geoengineering -All aspects of physics -E-governance -E-commerce -The knowledge economy -Urban studies -Arms control -Understanding and responding to terrorism -Medical informatics -Computer Sciences Applications Of Data Mining In Ebusiness And Finance, Frontiers in Artificial Intelligence and Applications, Volume 177 1 Cover 1 Frontiers in Artificial Intelligence and Applications Series 3 ISSN 0922-6389 3 Title Page 4 Editors: Carlos Soares, Yonghong Peng, Jun Meng, Takashi Washio, Zhi-Hua Zhou 4 Copyright 2008 The Authors and IOS Press 5 9781586038908 5 Preface 6 Program Committee 8 Contents 10 Applications of Data Mining in E-Business and Finance: Introduction 12 Preamble 12 Data Mining for Business 13 1. Application Issues in Data Mining 13 1.1. Business and Data Understanding 13 1.2. Data Preparation 14 1.3. Modeling 14 1.4. Evaluation 15 1.5. Deployment 15 2. Overview 16 2.1. Finance 17 2.2. E-Business 18 2.3. Other Applications 18 3. Conclusions 19 Acknowledgments 19 References 20 Evolutionary Optimization of Trading Strategies 22 Introduction 22 1. Problem Definition 23 1.1. Constrained Optimization Environment 24 2. Optimization with GA 26 2.1. Moving Average Strategies 26 2.2. Optimization with Genetic Algorithm 27 2.2.1. Fitness Function 27 2.2.2. Encoding and Search Space 27 2.2.3. Population Size 28 2.2.4. Crossover 28 2.2.5. Mutation 29 2.2.6. Evaluation History vs. Evaluation Time 29 2.3. Performance Boost 29 2.3.1. Data Storage 29 2.3.2. Parallel GA 30 3. Application and Re.nements 30 3.1. Empirical Studies in the Asx Market 31 3.2. Business Performance Stabilization 33 4. Conclusions 34 Acknowledgement 34 References 34 An Analysis of Support Vector Machines for Credit Risk Modeling 36 Introduction 36 1. Comparison of SVM and Logistic Regression 37 1.1. Experiments and Results 38 2. A Two Layer Cascade Model based on SVM 39 2.1. Idea and Analysis 39 2.2. The Cascade SVM-LR Model and Experiments 40 3. Probability of Default Modeling with SVM 41 4. Conclusion 43 References 43 A Tripartite Scorecard for the Pay/No pay Decision-Making in the Retail Banking Industry 56 Introduction 56 1. A Credit Model for the Pay/no pay Decision 57 1.1. Estimation of the loss matrix 57 2. Binary Scorecard 58 2.1. Tripartite Scorecard 59 3. Discussion 61 References 61 An Apriori Based Approach to Improve On-line Advertising Performance 62 Introduction 62 1. Related Work 64 2. Problem Statement and Our Approach 65 2.1. Apriori Rule Extraction 65 2.2. Rule Application 66 2.3. Performance Measurements 66 2.3.1. Banner-based Performance Measure 67 3. The Experiments 68 3.1. Data Collection 68 3.2. Results 68 4. Conclusions and Further Work 71 Acknowledgments 72 References 72 Probabilistic Latent Semantic Analysis for Search and Mining of Corporate Blogs 74 Introduction 74 1. Review of Related Work 76 1.1. Blog-speci.c Search Engines 76 1.2. Extraction of Useful Information from Blogs 76 2. Probabilistic Latent Semantic Analysis Model for Blog Mining 77 3. Experiments and Results 78 3.1. Data Set 78 3.2. Blog Search System 79 3.3. Results for Blog Mining of Topics 81 4. Conclusions 83 References 83 Towards Business Interestingness in Actionable Knowledge Discovery 110 Unknown -1 Faculty of Information Technology, University of Technology, Sydney, Australia e-mail: {dluo, lbcao, chaoluo, chengqi}@it.uts.ed 110 A2 Consulting Pty Limited, Sydney, Australia e-mail: {weiyuan.wang}@gmail.com 110 Introduction 110 business interestingness 110 objective 111 subjective 111 Pro.t 111 return 111 cost 111 bene.t 111 technical interestingness 111 objective 111 subjective business interestingness 111 actionable 111 1. Balancing Technical and Business Interestingness 112 1.1. Knowledge Actionability Studies Concentrating on Technical Signi.cance 112 1.2. Knowledge actionability satisfying technical signi.cance and business expectation 113 2. Case Study and Evaluation 114 3. Towards Domain-Driven, Actionable Knowledge Discovery 116 3.1. Aggregating Technical and Business Interestingness 116 Fuzzy weighting of individual interestingness measures. 117 Multi-objective optimization. 117 Fuzzy weighting of patterns. 117 3.2. Towards Domain Driven Data Mining 117 In-depth data intelligence. 118 Human intelligence. 118 Domain intelligence. 118 Web intelligence. 118 Intelligence meta-synthesis. 118 4. Conclusions 118 References 119 Sequence Mining for Business Analytics: Building Project Taxonomies for Resource Demand Forecasting 144 Introduction 144 1. Taxonomy Building as a Sequence Clustering Problem 146 2. HsMM-based Sequence Clustering 146 3. Resource Demand Forecasting based on Cluster Results 149 4. Results on IBM Data 150 5. Conclusions 152 References 152 Author Index 154 Applications Of Data Mining In Ebusiness And Finance,Frontiers in Artificial Intelligence,Apps,Vol 177 (2008) 157p 9781586038908 1586038907 Cover ......Page 1 ISSN 0922-6389......Page 3 Editors: Carlos Soares, Yonghong Peng, Jun Meng, Takashi Washio, Zhi-Hua Zhou......Page 4 9781586038908......Page 5 Preface......Page 6 Program Committee......Page 8 Contents......Page 10 Preamble......Page 12 1.1. Business and Data Understanding......Page 13 1.3. Modeling......Page 14 1.5. Deployment......Page 15 2. Overview......Page 16 2.1. Finance......Page 17 2.3. Other Applications......Page 18 Acknowledgments......Page 19 References......Page 20 Introduction......Page 22 1. Problem Definition......Page 23 1.1. Constrained Optimization Environment......Page 24 2.1. Moving Average Strategies......Page 26 2.2.2. Encoding and Search Space......Page 27 2.2.4. Crossover......Page 28 2.3.1. Data Storage......Page 29 3. Application and Re.nements......Page 30 3.1. Empirical Studies in the Asx Market......Page 31 3.2. Business Performance Stabilization......Page 33 References......Page 34 Introduction......Page 36 1. Comparison of SVM and Logistic Regression......Page 37 1.1. Experiments and Results......Page 38 2.1. Idea and Analysis......Page 39 2.2. The Cascade SVM-LR Model and Experiments......Page 40 3. Probability of Default Modeling with SVM......Page 41 References......Page 43 Introduction......Page 56 1.1. Estimation of the loss matrix......Page 57 2. Binary Scorecard......Page 58 2.1. Tripartite Scorecard......Page 59 References......Page 61 Introduction......Page 62 1. Related Work......Page 64 2.1. Apriori Rule Extraction......Page 65 2.3. Performance Measurements......Page 66 2.3.1. Banner-based Performance Measure......Page 67 3.2. Results......Page 68 4. Conclusions and Further Work......Page 71 References......Page 72 Introduction......Page 74 1.2. Extraction of Useful Information from Blogs......Page 76 2. Probabilistic Latent Semantic Analysis Model for Blog Mining......Page 77 3.1. Data Set......Page 78 3.2. Blog Search System......Page 79 3.3. Results for Blog Mining of Topics......Page 81 References......Page 83 business interestingness......Page 110 Unknown......Page 0 actionable......Page 111 1.1. Knowledge Actionability Studies Concentrating on Technical Signi.cance......Page 112 1.2. Knowledge actionability satisfying technical signi.cance and business expectation......Page 113 2. Case Study and Evaluation......Page 114 3.1. Aggregating Technical and Business Interestingness......Page 116 3.2. Towards Domain Driven Data Mining......Page 117 4. Conclusions......Page 118 References......Page 119 Introduction......Page 144 2. HsMM-based Sequence Clustering......Page 146 3. Resource Demand Forecasting based on Cluster Results......Page 149 4. Results on IBM Data......Page 150 References......Page 152 Author Index......Page 154 The application of Data Mining (DM) technologies has shown an explosive growth in an increasing number of different areas of business, government and science. Two of the most important business areas are finance, in particular in banks and insurance companies, and e-business, such as web portals, e-commerce and ad management services.In spite of the close relationship between research and practice in Data Mining, it is not easy to find information on some of the most important issues involved in real world application of DM technology, from business and data understanding to evaluation and deployment. Papers often describe research that was developed without taking into account constraints imposed by the motivating application. When these issues are taken into account, they are frequently not discussed in detail because the paper must focus on the method. Therefore knowledge that could be useful for those who would like to apply the same approach on a related problem is not shared. The papers in this book address some of these issues. This book is of interest not only to Data Mining researchers and practitioners, but also to students who wish to have an idea of the practical issues involved in Data Mining. IOS Press is an international science, technical and medical publisher of high-quality books for academics, scientists, and professionals in all fields. Some of the areas we publish -Biomedicine -Oncology -Artificial intelligence -Databases and information systems -Maritime engineering -Nanotechnology -Geoengineering -All aspects of physics -E-governance -E-commerce -The knowledge economy -Urban studies -Arms control -Understanding and responding to terrorism -Medical informatics -Computer Sciences In spite of the close relationship between research and practice in Data Mining, it is not easy to find information on some of the important issues involved in real world application of DM technology. This book address some of these issues. It is suitable for Data Mining researchers and practitioners.