Visual Data Mining: Theory, Techniques and Tools for Visual Analytics (Lecture Notes in Computer Science)

Annotation The importance of visual data mining, as a strong sub-discipline of data mining, had already been recognized in the beginning of the decade. In 2005 a panel of renowned individuals met to address the shortcomings and drawbacks of the current state of visual information processing. The need for a systematic and methodological development of visual analytics was detected. This book aims at addressing this need. Through a collection of 21 contributions selected from more than 46 submissions, it offers a systematic presentation of the state of the art in the field. The volume is structured in three parts on theory and methodologies, techniques, and tools and applications Title Page Foreword Preface Table of Contents Visual Data Mining: An Introduction and Overview Introduction The Term The Process The Book Conclusions and Acknowledgements References The 3DVDM Approach: A Case Study with Clickstream Data Introduction Clickstreams Web Server Log Files Modeling Clickstreams in a DataWarehouse Analyzing Clickstreams The 3DVDM System Overall Architecture Streaming Visible Data Model Computation Data Analyzes and Interpretation Animation Conditional Density Surfaces Equalization of Structures Windowing Summary and Future Work References Form-Semantics-Function – A Framework for Designing Visual Data Representations for Visual Data Mining Introduction Form-Semantics-Function: A Formal Approach Towards Constructing and Evaluating Visualisation Techniques Metaphor Analysis Metaphor Formalisation Metaphor Evaluation Constructing Visualisation Schema for Visual Data Mining for Identifying Patterns in Team Collaboration Metaphor Analysis Metaphor Formalisation Evaluation and Comparison of Two Visualisation Schemata Metaphor Evaluation Conclusion and Future Directions References A Methodology for Exploring Association Models Introduction Association Rules The Association Rule Space PEAR: A Web-Based AR Browser Chunking Large Sets of Rules Global Metrics for Sets of Rules The Index Page Operators for Sets of Association Rules Example of the Application of the Proposed Methodology Implementation Scalable Vector Graphics Representing Associations Rules with PMML Performance Related Work Future Work and Conclusions References Visual Exploration of Frequent Itemsets and Association Rules Introduction BasicConcepts Itemset Lattice and Closure Properties Parallel Coordinates Dealing with Item Taxonomy Experiments and Screen Snapshots Visualization of Iceberg Data Cubes Related Work Conclusion and Future Work References Visual Analytics: Scope and Challenges Introduction Scope of Visual Analytics Visual Analytics Process Application Challenges Physics and Astronomy Business Environmental Monitoring Disaster and Emergency Management Security Software Analytics Biology, Medicine and Health Engineering Analytics Personal Information Management Mobile Graphics and Traffic Technical Challenges Conclusion References Using Nested Surfaces for Visual Detection of Structures in Databases Introduction Motivation Preliminaries Probability Density Function Clusters and Outliers Surface Definition Algorithms Evaluation Quality of the Surfaces Space and Time Complexities Experiments Summary and Future Work References Visual Mining of Association Rules Introduction Some Issues about Association Rules A Real Data Set Application Visualizing Association Rules Rule Table Two-Dimensional Matrix 3-D Visualization Association Rules Networks The TwoKey Plot Double-Decker Plot Parallel Coordinates Factorial Planes Concluding Remarks References Interactive Decision Tree Construction for Interval and Taxonomical Data Introduction Interactive Decision Tree Construction Interval Data Ordering Interval Data Graphical Representation of Interval Data Classifying Interval Data with PBC Classifying Interval Data with CIAD Interval SVM Algorithm Taxonomical Data Graphical Representation of a Taxonomical Variable Interactive Taxonomical Data Classification Some Results Conclusion and Future Work References Visual Methods for Examining SVM Classifiers Introduction Methods Support Vector Machines Tours Methods for Visualization SVM and Tours Application Data Description Visualizing SVM Outputs Gene Selection Varying SVM Input Parameters Model Stability Summary and Discussion Summary Discussion References Text Visualization for Visual Text Analytics Introduction Overview of Visual Text Analytics Technology Functional Components of Visual Text Analytics Systems Tokenization Vector Representation Dimensionality Reduction Spatialization Labeling Interactive Exploration Summary Example Systems Sammon Lin Bead IN-SPIRE WEBSOM Starlight References Visual Discovery of Network Patterns of Interaction between Attributes Introduction Modeling Perspectives in Analytics Network Models in Analytics The “Loss of Detail” Problem in Data Mining The “Independency of Attributes” Assumption in Data Mining Visual Discovery of Network Patterns of Interaction between Attributes The Approach and the Processes Case Studies Case A: Fraud Detection in Insurance Industry Case B: Visual Discovery in Internet Traffic Analysis Conclusion References Mining Patterns for Visual Interpretation in a Multiple-Views Environment Introduction Related Work Multiple-Views within the Visualization Tree Visualization Pipeline Visualization Composition The System Features of the VisTree Methodology Exploration Techniques Frequency Plot Relevance Plot Representative Plot Experiments Conclusions References Using 2D Hierarchical Heavy Hitters to Investigate Binary Relationships Introduction Related Work Two-Dimensional Hierarchical Heavy Hitters Filtering Grouping 1D Hierarchical Data Grouping 2D Hierarchical Data Computation of Counts in a Lattice VHHH: Visualization of Hierarchical Heavy Hitters Visualization of the Lattice and HHH Information Ordering of Categorical Data Experiments Case Studies with Real World Data VHHH Alphabet Pattern Investigation of VHHH HHH Ordering Versus Dataset Ordering Conclusions and Future Work References Complementing Visual Data Mining with the Sound Dimension: Sonification of Time Dependent Data Introduction Characteristics of Sound for Time Dependent Data Representation Sonification Detection of Outliers Beat Drums Mapping Stereo Panning 3D Curve Experimental Workbench for Data Sonification and Mining Design of the Experiment and Methodology of Data Collection Results of the Experiments The Sample of Participants Results in 2D Results in 3D Discussion Conclusions References Context Visualization for Visual Data Mining Introduction Formal Model of Interactive Visualization for Visual Data Mining Interactive Navigation in Information Visualization Visual Exploration in Visual Data Mining The Concept of Visual Exploration with a Chain of Context Views Context Visualization with a Chain of Context Views History Visualization for Visual Data Mining Conclusion References Assisting Human Cognition in Visual Data Mining Introduction Visual Bias in Visual Data Mining Addressing the Visual Bias in Visual Data Mining The Method of Guided Cognition, Implemented through Embedded Statistical Techniques The Method of Validated Cognition, Implemented through a Combination of Visual Data Mining Techniques Visual Analysis Validation Summary and Future Directions References Immersive Visual Data Mining: The 3DVDM Approach Introduction Virtual Reality Immersive Visual Data Mining Exploiting Sound Previous Work Visual Data Exploration Auditory Data Exploration Immersive vs. Traditional VDM The3DVDMSystem VR++ and 3DVDM Data Pipeline and Interaction Principles of Data Visualization in 3DVDM Rendering Sound BasicTools Visual Data Exploration Auditory Data Exploration Methodology for Visual Data Exploration in 3D Worlds Data Preparation Basic Statistical Analysis Visual Exploration of Static Worlds 3D Scatter Plot Tour 3D Scatter Plots and Object Properties Visual Exploration of Dynamic Worlds Macro Dynamic Visualization Micro Dynamic Visualization Auditory Exploration of Static Worlds Sound Supporting Color Sound “On Its Own” – Categorical Variables Sound “On Its Own” – Continuous Variables Discussion Visual Data Exploration Auditory Data Exploration Conclusions References DataJewel: Integrating Visualization with Temporal Data Mining Introduction Related Work User-Centric Data Mining The Visualization Component CalendarView Interaction with CalendarView The Temporal Mining Component The Database Component Experiments Mining Airplane Maintenance Datasets The DataJewel System Discussion Conclusions References A Visual Data Mining Environment Introduction Related Work System Architecture and Implementation System Architecture System Implementation The User Interface Identifying Regions with Good Sales: Using the Clustering Environment Establishing Data Relationships: Using the Metaquery Environment Market-Basket Analysis: Using the Association Rule Environment Visual Exploration Using DARE Formal Specification of the Visual Interface Clustering Usability Heuristic Evaluation Mock-Up Experiment User Tests Future Work and Conclusions References Integrative Visual Data Mining of Biomedical Data: Investigating Cases in Chronic Fatigue Syndrome and Acute Lymphoblastic Leukaemia Introduction The “Extract-Explain-Generate” Methodology Case Study 1: Chronic Fatigue Syndrome Problem The Goals of the Study The Study Scenario The Outcomes Case Study 2: Acute Lymphoblastic Leukaemia Problem The Goals of the Study The Study Scenario The Outcomes Discussion and Conclusions References Towards Effective Visual Data Mining with Cooperative Approaches Introduction Interactive Decision Tree Construction CIAD CIAD+ Some Results of Interactive Decision Tree Algorithms Visualization of SVM Results Visualization of the SVM Separating Plane Visualization of the Data Distribution According to the Distance to the Boundary Visualization to Tune SVM Input Parameters Cooperative Approaches for Large Datasets Interactive SVM Construction Cooperative Approach for Datasets with Large Number of Datapoints Cooperative Approach for Datasets with Large Number of Dimensions Conclusion and Future Work References Author Index Simeon J. Simoff, Michael H. Böhlen, Arturas Mazeika (eds.). Includes Bibliographical References And Index. Also Available Via The Internet.