Data Mining for Co-location Patterns: Principles and Applications

Cover Page
Half Title Page
Title Page
Copyright Page
Contents Page
Preface Page
Acknowledgments Page
Author Biography Page
Chapter 1 Introduction
	1.1 Background
	1.2 Data Mining
		1.2.1 Concept for Data Mining
		1.2.2 Data Mining and Knowledge Discovery
		1.2.3 Data Mining with Other Disciplines
		1.2.4 Data Mining Tasks
	1.3 Geospatial Data Mining
	1.4 Comparison between Spatial Data Mining and Data Mining
		1.4.1 Spatial Data in Data Mining
		1.4.2 Data Mining and Spatial Database
	1.5 Decision Trees and Decision Rules
		1.5.1 Decision Tree Induction
		1.5.2 Decision Tree Modeling
			1.5.2.1 Growth Phase
			1.5.2.2 Pruning Phase
		1.5.3 Measures for Selecting the Best Split
			1.5.3.1 Entropy
			1.5.3.2 Information Gain
			1.5.3.3 Gain Information Ratio
		1.5.4 Decision Rule Induction
		1.5.5 Evaluation of the Performance of Decision Tree
			1.5.5.1 Accuracy of Performance
			1.5.5.2 Twofold Cross-Validation
		1.5.6 Problems of Decision Tree Induction Data Mining
	1.6 Co-Location Pattern Mining
	1.7 Arrangement of the Chapters
	References
Chapter 2 Fundamentals of Mining Co-Location Patterns
	2.1 Basic Concepts of Mining Co-Location Patterns
	2.2 Three Basic Types of Co-Location Pattern Mining Algorithms
		2.2.1 Join-Based Algorithms
		2.2.2 Partial Join Algorithms
		2.2.3 Join-Less Algorithms
		2.2.4 Advantages and Disadvantages of Three Basic Algorithms
		2.2.5 Other Algorithms
			2.2.5.1 Co-Location Pattern Mining Algorithms with Rare Spatial Features
			2.2.5.2 Maximal Clique Algorithms
			2.2.5.3 Density Based Co-Location Pattern Mining Algorithms
			2.2.5.4 Co-Location Pattern Mining Algorithms with Fuzzy Attributes
			2.2.5.5 Co-Location Pattern Mining Algorithms with Time Constraints
	2.3 Spatial Negative Co-Location Mining Algorithms
	2.4 Differences between Positive and Negative Co-Location Pattern Mining
	2.5 Summary of This Chapter
	References
Chapter 3 Principle of Mining Co-Location Patterns
	3.1 Introduction
	3.2 Co-Location Mining Algorithms
		3.2.1 Definitions of the Co-Location Mining Method
		3.2.2 Principles of Co-Location Pattern Mining Algorithms
			3.2.2.1 Initialization
			3.2.2.2 Determination of Candidate Co-Locations
			3.2.2.3 Determination of Table Instances of Candidate Co-Locations
			3.2.2.4 Pruning
			3.2.2.5 Generating Co-Location Rules
	3.3 Co-Location Decision Tree (CL-DT) Algorithms
		3.3.1 CL-DT Algorithm Modeling
		3.3.2 Attribute Selection
		3.3.3 Co-Location Mining Rules
		3.3.4 Node Merging Criteria
		3.3.5 Decision Rule Induction from CL-DT
	3.4 Linear Multivariate CL-DT Algorithms
	3.5 Example Analysis
		3.5.1 Decision Tree and Decision Rules Induction Using a C4.5 Algorithm
		3.5.2 CL-DT Algorithms
			3.5.2.1 CL-DT Mining Rules
			3.5.2.2 CL-DT Induction
	3.6 Discussion and Analysis of CL-DT
	References
Chapter 4 Manifold Learning Co-Location Pattern Mining
	4.1 Introduction
	4.2 MVU-Based Co-Location Pattern Mining
		4.2.1 Brief Review of MVU
		4.2.2 MVU-Based Co-Location Pattern Mining
		4.2.3 MVU-Based Co-Location Mining Rules
			4.2.3.1 MVU Unfolded Distance Algorithms
				4.2.3.1.1 Neighbor Relation Matrix Reservation
				4.2.3.1.2 MVU Function Establishment and Solution
				4.2.3.1.3 The Calculation of the Unfolded Distance Between Instances
			4.2.3.2 Determination of MVU-Based Co-Locations Patterns
			4.2.3.3 Determination of Distinct Event-Types
		4.2.4 Generation of MVU-Based Co-Location Rules
	4.3 Pruning
	4.4 Inducting Decision Rules
	References
Chapter 5 Maximal Instance Co-Location Pattern Mining Algorithms
	5.1 Introduction
	5.2 Maximal Instance Algorithms
		5.2.1 Generation of Row Instances
	5.3 RI-Tree Construction
		5.3.1 Rules of RI-Tree
		5.3.2 Completeness of RI-Tree
	5.4 Generation of Co-Locations
	5.5 Discussions for Maximal Instance Algorithms
		5.5.1 Comparison Analysis of Row Instance Generation
		5.5.2 Comparison Analysis of Maximal Instance Algorithms
	References
Chapter 6 Negative Co-Location Pattern Mining Algorithms
	6.1 Introduction
	6.2 Definition and Lemma for Negative Co-Location
		6.2.1 Basic Definition of Negative Co-Location
		6.2.2 Lemmas for Negative Co-Location
	6.3 Algorithm for Mining Negative Co-Location Patterns
		6.3.1 Generation of Candidate Negative Co-Locations
		6.3.2 Pruning
	6.4 Join-Based Prevalent Negative Co-Location Patterns
	6.5 Experiment and Analysis
		6.5.1 Data Sets
		6.5.2 Join-Based Prevalent Negative Co-Location Patterns
		6.5.3 Difficulties in Mining Negative Co-Location Patterns
	6.6 Conclusions
	References
Chapter 7 Application of Mining Co-Location Patterns in Pavement Management and Rehabilitation
	7.1 Introduction
		7.1.1 Distress Rating
		7.1.2 Potential Rehabilitation Strategies
	7.2 Experimental Design
		7.2.1 Flowchart of Experiment and Comparison Analysis
		7.2.2 Data Sources
		7.2.3 Nonspatial Attribute Data Selection
		7.2.4 Spatial Attribute Data Selection
		7.2.5 Maintenance and Rehabilitation (M&R) Strategies
	7.3 Induction of Co-Location Mining Rules
		7.3.1 Determination of Candidate Co-Locations
		7.3.2 Determination of Table Instances of Candidate Co-Locations
			7.3.2.1 Determination of Distinct Events
			7.3.2.2 Co-Location Mining for Rehabilitation and Maintenance Strategy
				7.3.2.2.1 Crack Pouring (CP)
				7.3.2.2.2 Full-Depth Patch (FDP)
				7.3.2.2.3 1” Plant Mix Resurfacing (PM1)
				7.3.2.2.4 2” Plant Mix Resurfacing (PM2)
				7.3.2.2.5 Skin Patch (SKP)
				7.3.2.2.6 Short Overlay (SO)
			7.3.2.3 Pruning
		7.3.3 Generating Co-Location Rules
	7.4 Experiments of CL-DT Induction
		7.4.1 Basic Steps of CL-DT Induction
			7.4.1.1 Step 1. Load Pavement Database
			7.4.1.2 Step 2. Data Inputs
		7.4.2 Experimental Results
			7.4.2.1 Induced Decision Tree
			7.4.2.2 Induced Decision Rules
	7.5 Mapping of CL-DT-Based Decision of M&R
	7.6 Comparison Analysis and Discussion
		7.6.1 Comparison Analysis for the Induced Decision Tree Parameter
		7.6.2 Comparison Analysis for the Misclassified Percentage
		7.6.3 Comparison Analysis for the Computational Time
		7.6.4 Comparison Analysis of Support, Confidence, and Capture for Rule Induction
		7.6.5 Verification of the Quantity of Each Treatment Strategy
		7.6.6 Verification of the Location of Each Treatment Strategy
	7.7 Discussion and Remarks for Co-Location Decision Tree Algorithm
	7.8 Conclusions
		7.8.1 Advantages and Disadvantages of Applying Existing DT Method in Pavement M&R Strategy Decision Making
		7.8.2 Significances of the Proposed CL-DT Method for Pavement M&R Strategy Decision Making
	References
Chapter 8 Application of Mining Co-Location Patterns in Buffer Analysis
	8.1 Introduction
	8.2 Generalized Buffering Algorithms
	8.3 Discussion of Three Types of GBAS
		8.3.1 Generalized Point Buffering Algorithms
		8.3.2 Generalized Line Buffering
		8.3.3 Generalized Polygon Buffering
	8.4 Experiments and Analysis
		8.4.1 Data Sets
		8.4.2 Generalized Buffer Analysis
			8.4.2.1 Experiment for Data Set-1
				8.4.2.1.1 Generalized Point Buffer Analysis
				8.4.2.1.2 Generalized Line Buffer Algorithms
				8.4.2.1.3 Generalized Polygon Buffer (GPLB) Generation
			8.4.2.2 Experiment for Data Set-2
				8.4.2.2.1 Generalized Point Buffer Zone Generation
				8.4.2.2.2 Generalized Line Buffer Analysis
				8.4.2.2.3 Generalized Polygon Buffer Analysis
		8.4.3 Comparison Analysis and Remarks
			8.4.3.1 Comparison Analysis
			8.4.3.2 Remarks from the Compared Results
				8.4.3.2.1 From Traditional/Generalized Point Buffer Algorithms
				8.4.3.2.2 From Traditional/Generalized Line Buffer Algorithms
				8.4.3.2.3 From Traditional/Generalized Polygon Buffer Algorithms
	8.5 Conclusion
	References
Chapter 9 Application of Mining Co-Location Patterns in Remotely Sensed Imagery Classification
	9.1 Introduction
	9.2 Data Sets
		9.2.1 Data Sets
		9.2.2 Nonspatial Attribute and Spatial Attribute Selection
	9.3 Experiments
		9.3.1 Experiments on the First Test Area
			9.3.1.1 Input Parameters
			9.3.1.2 Generation of MVU-Based Co-Location Mining Rules
				9.3.1.2.1 Calculation of Unfolded Distances
				9.3.1.2.2 Determination of the MVU-Based Co-Location
				9.3.1.2.3 Determination of Distinct-Type Events
			9.3.1.3 Experimental Results
		9.3.2 Experiments on the Second Test Area
	9.4 Comparison Analysis and Validation in the Field
		9.4.1 Classification Accuracy Comparison
		9.4.2 Parameters and Computation Time Comparison
			9.4.2.1 Comparison of the Induced Decision Tree Parameters
			9.4.2.2 Comparison of the Computational Time
		9.4.3 Validation in Field
	9.5 Conclusions
	References
Index