Data Mining for Business Applications

by Cao, Longbing; Yu, Philip S.; Zhang, Chengqi; Zhang, Huaifeng

• ISBN13: 9780387794198
• ISBN10: 0387794190
• Format: Hardcover
• Copyright: 2008-11-01
• Publisher: Springer-Verlag New York Inc

Summary

"Data Mining for Business Applications presents the state-of-the-art research and development outcomes on methodologies, techniques, approaches and successful applications in the area. The contributions mark a paradigm shift from "data-centered pattern mining" to "domain driven actionable knowledge discovery" for next-generation KDD research and applications. The contents identify how KDD techniques can better contribute to critical domain problems in theory and practice, and strengthen business intelligence in complex enterprise applications. The volume also explores challenges and directions for future research and development in the dialogue between academia and business."--BOOK JACKET.

Table of Contents

Domain Driven KDD Methodology
Introduction to Domain Driven Data Mining	p. 3
Why Domain Driven Data Mining	p. 3
What Is Domain Driven Data Mining	p. 5
Basic Ideas	p. 5
D3M for Actionable Knowledge Discovery	p. 6
Open Issues and Prospects	p. 9
Conclusions	p. 9
References	p. 10
Post-processing Data Mining Models for Actionability	p. 11
Introduction	p. 11
Plan Mining for Class Transformation	p. 12
Overview of Plan Mining	p. 12
Problem Formulation	p. 14
From Association Rules to State Spaces	p. 14
Algorithm for Plan Mining	p. 17
Summary	p. 19
Extracting Actions from Decision Trees	p. 20
Overview	p. 20
Generating Actions from Decision Trees	p. 22
The Limited Resources Case	p. 23
Learning Relational Action Models from Frequent Action Sequences	p. 25
Overview	p. 25
ARMS Algorithm: From Association Rules to Actions	p. 26
Summary of ARMS	p. 28
Conclusions and Future Work	p. 29
References	p. 29
On Mining Maximal Pattern-Based Clusters	p. 31
Introduction	p. 32
Problem Definition and Related Work	p. 34
Pattern-Based Clustering	p. 34
Maximal Pattern-Based Clustering	p. 35
Related Work	p. 35
Algorithms MaPle and MaPle+	p. 36
An Overview of MaPle	p. 37
Computing and Pruning MDS's	p. 38
Progressively Refining, Depth-first Search of Maximal pClusters	p. 40
MaPle+: Further Improvements	p. 44
Empirical Evaluation	p. 46
The Data Sets	p. 46
Results on Yeast Data Set	p. 47
Results on Synthetic Data Sets	p. 48
Conclusions	p. 50
References	p. 50
Role of Human Intelligence in Domain Driven Data Mining	p. 53
Introduction	p. 53
DDDM Tasks Requiring Human Intelligence	p. 54
Formulating Business Objectives	p. 54
Setting up Business Success Criteria	p. 55
Translating Business Objective to Data Mining Objectives	p. 56
Setting up of Data Mining Success Criteria	p. 56
Assessing Similarity Between Business Objectives of New and Past Projects	p. 57
Formulating Business, Legal and Financial Requirements	p. 57
Narrowing down Data and Creating Derived Attributes	p. 58
Estimating Cost of Data Collection, Implementation and Operating Costs	p. 58
Selection of Modeling Techniques	p. 59
Setting up Model Parameters	p. 59
Assessing Modeling Results	p. 59
Developing a Project Plan	p. 60
Directions for Future Research	p. 60
Summary	p. 61
References	p. 61
Ontology Mining for Personalized Search	p. 63
Introduction	p. 63
Related Work	p. 64
Architecture	p. 65
Background Definitions	p. 66
World Knowledge Ontology	p. 66
Local Instance Repository	p. 67
Specifying Knowledge in an Ontology	p. 68
Discovery of Useful Knowledge in LIRs	p. 70
Experiments	p. 71
Experiment Design	p. 71
Other Experiment Settings	p. 74
Results and Discussions	p. 75
Conclusions	p. 77
References	p. 77
Novel KDD Domains & Techniques
Data Mining Applications in Social Security	p. 81
Introduction and Background	p. 81
Case Study I: Discovering Debtor Demographic Patterns with Decision Tree and Association Rules	p. 83
Business Problem and Data	p. 83
Discovering Demographic Patterns of Debtors	p. 83
Case Study II: Sequential Pattern Mining to Find Activity Sequences of Debt Occurrence	p. 85
Impact-Targeted Activity Sequences	p. 86
Experimental Results	p. 87
Case Study III: Combining Association Rules from Heterogeneous Data Sources to Discover Repayment Patterns	p. 89
Business Problem and Data	p. 89
Mining Combined Association Rules	p. 89
Experimental Results	p. 90
Case Study IV: Using Clustering and Analysis of Variance to Verify the Effectiveness of a New Policy	p. 92
Clustering Declarations with Contour and Clustering	p. 92
Analysis of Variance	p. 94
Conclusions and Discussion	p. 94
References	p. 95
Security Data Mining: A Survey Introducing Tamper-Resistance	p. 97
Introduction	p. 97
Security Data Mining	p. 98
Definitions	p. 98
Specific Issues	p. 99
General Issues	p. 101
Tamper-Resistance	p. 102
Reliable Data	p. 102
Anomaly Detection Algorithms	p. 104
Privacy and Confidentiality Preserving Results	p. 105
Conclusion	p. 108
References	p. 108
A Domain Driven Mining Algorithm on Gene Sequence Clustering	p. 111
Introduction	p. 111
Related Work	p. 112
The Similarity Based on Biological Domain Knowledge	p. 114
Problem Statement	p. 114
A Domain-Driven Gene Sequence Clustering Algorithm	p. 117
Experiments and Performance Study	p. 121
Conclusion and Future Work	p. 124
References	p. 125
Domain Driven Tree Mining of Semi-structured Mental Health Information	p. 127
Introduction	p. 127
Information Use and Management within Mental Health Domain	p. 128
Tree Mining - General Considerations	p. 130
Basic Tree Mining Concepts	p. 131
Tree Mining of Medical Data	p. 135
Illustration of the Approach	p. 139
Conclusion and Future Work	p. 139
References	p. 140
Text Mining for Real-time Ontology Evolution	p. 143
Introduction	p. 144
Related Text Mining Work	p. 145
Terminology and Multi-representations	p. 145
Master Aliases Table and OCOE Data Structures	p. 149
Experimental Results	p. 152
CAV Construction and Information Ranking	p. 153
Real-Time CAV Expansion Supported by Text Mining	p. 154
Conclusion	p. 155
Acknowledgement	p. 156
References	p. 156
Microarray Data Mining: Selecting Trustworthy Genes with Gene Feature Ranking	p. 159
Introduction	p. 159
Gene Feature Ranking	p. 161
Use of Attributes and Data Samples in Gene Feature Ranking	p. 162
Gene Feature Ranking: Feature Selection Phase 1	p. 163
Gene Feature Ranking: Feature Selection Phase 2	p. 163
Application of Gene Feature Ranking to Acute Lymphoblastic Leukemia data	p. 164
Conclusion	p. 166
References	p. 167
Blog Data Mining for Cyber Security Threats	p. 169
Introduction	p. 169
Review of Related Work	p. 170
Intelligence Analysis	p. 171
Information Extraction from Blogs	p. 171
Probabilistic Techniques for Blog Data Mining	p. 172
Attributes of Blog Documents	p. 172
Latent Dirichlet Allocation	p. 173
Isometric Feature Mapping (Isomap)	p. 174
Experiments and Results	p. 175
Data Corpus	p. 175
Results for Blog Topic Analysis	p. 176
Blog Content Visualization	p. 178
Blog Time Visualization	p. 179
Conclusions	p. 180
References	p. 181
Blog Data Mining: The Predictive Power of Sentiments	p. 183
Introduction	p. 183
Related Work	p. 185
Characteristics of Online Discussions	p. 186
Blog Mentions	p. 186
Box Office Data and User Rating	p. 187
Discussion	p. 187
S-PLSA: A Probabilistic Approach to Sentiment Mining	p. 188
Feature Selection	p. 188
Sentiment PLSA	p. 188
ARSA: A Sentiment-Aware Model	p. 189
The Autoregressive Model	p. 190
Incorporating Sentiments	p. 191
Experiments	p. 192
Experiment Settings	p. 192
Parameter Selection	p. 193
Conclusions and Future Work	p. 194
References	p. 194
Web Mining: Extracting Knowledge from the World Wide Web	p. 197
Overview of Web Mining Techniques	p. 197
Web Content Mining	p. 199
Classification: Multi-hierarchy Text Classification	p. 199
Clustering Analysis: Clustering Algorithm Based on Swarm Intelligence and k-Means	p. 200
Semantic Text Analysis: Conceptual Semantic Space	p. 202
Web Structure Mining: Page Rank vs. HITS	p. 203
Web Event Mining	p. 204
Preprocessing for Web Event Mining	p. 205
Multi-document Summarization: A Way to Demonstrate Event's Cause and Effect	p. 206
Conclusions and Future Works	p. 206
References	p. 207
DAG Mining for Code Compaction	p. 209
Introduction	p. 209
Related Work	p. 211
Graph and DAG Mining Basics	p. 211
Graph-based versus Embedding-based Mining	p. 212
Embedded versus Induced Fragments	p. 213
DAG Mining Is NP-complete	p. 213
Algorithmic Details of DAGMA	p. 214
A Canonical Form for DAG enumeration	p. 214
Basic Structure of the DAG Mining Algorithm	p. 215
Expansion Rules	p. 216
Application to Procedural Abstraction	p. 219
Evaluation	p. 220
Conclusion and Future Work	p. 222
References	p. 223
A Framework for Context-Aware Trajectory Data Mining	p. 225
Introduction	p. 225
Basic Concepts	p. 227
A Domain-driven Framework for Trajectory Data Mining	p. 229
Case Study	p. 232
The Selected Mobile Movement-aware Outdoor Game	p. 233
Transportation Application	p. 234
Conclusions and Future Trends	p. 238
References	p. 239
Census Data Mining for Land Use Classification	p. 241
Content Structure	p. 241
Key Research Issues	p. 242
Land Use and Remote Sensing	p. 242
Census Data and Land Use Distribution	p. 243
Census Data Warehouse and Spatial Data Mining	p. 243
Concerning about Data Quality	p. 243
Concerning about Domain Driven	p. 244
Applying Machine Learning Tools	p. 246
Data Integration	p. 247
Area of Study and Data	p. 247
Supported Digital Image Processing	p. 248
Putting All Steps Together	p. 248
Results and Analysis	p. 249
References	p. 251
Visual Data Mining for Developing Competitive Strategies in Higher Education	p. 253
Introduction	p. 253
Square Tiles Visualization	p. 255
Related Work	p. 256
Mathematical Model	p. 257
Framework and Case Study	p. 260
General Insights and Observations	p. 261
Benchmarking	p. 262
High School Relationship Management (HSRM)	p. 263
Future Work	p. 264
Conclusions	p. 264
References	p. 265
Data Mining For Robust Flight Scheduling	p. 267
Introduction	p. 267
Flight Scheduling in the Presence of Delays	p. 268
Related Work	p. 270
Classification of Flights	p. 272
Subspaces for Locally Varying Relevance	p. 272
Integrating Subspace Information for Robust Flight Classification	p. 272
Algorithmic Concept	p. 274
Monotonicity Properties of Relevant Attribute Subspaces	p. 274
Top-down Class Entropy Algorithm: Lossless Pruning Theorem	p. 275
Algorithm: Subspaces, Clusters, Subspace Classification	p. 276
Evaluation of Flight Delay Classification in Practice	p. 278
Conclusion	p. 280
References	p. 280
Data Mining for Algorithmic Asset Management	p. 283
Introduction	p. 283
Backbone of the Asset Management System	p. 285
Expert-based Incremental Learning	p. 286
An Application to the iShare Index Fund	p. 290
References	p. 294
Reviewer List	p. 297
Index	p. 299
Table of Contents provided by Publisher. All Rights Reserved.