Intelligent Decision Support Systems

✍ Scribed by Miquel Sànchez-Marrè

• Publisher: Springer
• Year: 2022
• Tongue: English
• Leaves: 826
• Category: Library

✦ Synopsis

This book, with invaluable contributions of Professor Franz Wotawa in chapters 5 and 7, presents the potential use and implementation of intelligent techniques in decision making processes involved in organizations and companies. It provides a thorough analysis of decisions, reviewing the classical decision theory, and describing usual methods for modeling the decision process. It describes the chronological evolution of Decision Support Systems (DSS) from early Management Information Systems until the appearance of Intelligent Decision Support Systems (IDSS). It explains the most commonly used intelligent techniques, both data-driven and model-driven, and illustrates the use of knowledge models in Decision Support through case studies. The author pays special attention to the whole Data Science process, which provides intelligent data-driven models in IDSS. The book describes main uncertainty models used in Artificial Intelligence to model inexactness; covers recommender systems; and reviews available development tools for inducing data-driven models, for using model-driven methods and for aiding the development of Intelligent Decision Support Systems.

✦ Table of Contents

Foreword
Preface
To Whom
Structure of the Book
Contributor
Acknowledgements
Contents
Author and Contributor
Abbreviations
Part I: Fundamentals
Chapter 1: Introduction
1.1 Complexity of Real-World System
1.2 The Need of Decision Support Tools
References
Further Reading
Chapter 2: Decisions
2.1 Fundamentals About Decisions
2.2 Decision Typologies
2.3 Decision Theory
2.3.1 Origins of Decision Theory
2.3.2 Modern Decision Theory
2.3.2.1 Analyzing the Decision Process
2.3.2.1.0 Sequential Models
2.3.2.1.0 Non-sequential Model
2.3.2.2 Managing the Decision Process
2.3.2.2.0 Facing the Intelligence Phase
2.3.2.2.0 Facing the Design Phase
2.3.2.2.0 Facing the Choice Phase
2.4 Decision Process Modelling
2.4.1 Single Decision Scenario
2.4.1.1 Decision-Making Under Certainty
2.4.1.1.0 Single-Attribute Approach
2.4.1.1.0 Multiple-Attribute Approach
2.4.1.2 Decision-Making Under Risk
2.4.1.3 Decision-Making Under Uncertainty
2.4.1.3.0 Decision-Making Under Classical Ignorance
2.4.1.3.0 Decision-Making Under Unknown Consequences
2.4.1.4 Decision-Making Under Hybrid Scenarios
2.4.2 Multiple Sequential Decisions
2.4.2.1 Decision Trees
2.4.2.2 Influence Diagrams
2.5 Group Decision-Making
References
Further Reading
Chapter 3: Evolution of Decision Support Systems
3.1 Historical Perspective of Management Information Systems
3.2 Decision Support Systems
3.3 Classification of DSS
3.3.1 Model-Driven DSSs
3.3.1.1 First-Principles Models or Mechanistic Models
3.3.1.2 Decision Analysis Tools and Models
3.3.1.2.0 Analytic Hierarchy Process Model
3.3.1.2.0 Decision Matrix and Decision Table models
3.3.1.2.0 Decision Tree Model
3.3.1.3 Optimization Models
3.3.1.3.0 Multi-Criteria Decision Analysis (MCDA) Models
3.3.1.4 Simulation Models
3.3.1.4.0 Monte Carlo Simulation Models
3.3.1.4.0 Discrete-Event Simulation Models
3.3.2 Data-Driven DSSs
3.3.2.1 Direct Data Exploration DSSs
3.3.2.1.0 Data-Reporting DSSs
3.3.2.1.0 Data-Analytic or Data-Intensive DSSs
3.3.2.1.0 Database Querying Model DSSs
3.3.2.1.0 Executive Information System (EIS) Model DSSs
3.3.2.1.0 Data Warehouse and OLAP System Model DSSs
3.3.2.2 Data Model Exploration DSSs
3.3.2.2.0 Data Confirmative Model DSSs
3.3.2.2.0 Data Explorative Model DSSs
3.4 The Interpretation Process in Decision Support Systems
References
Further Reading
Part II: Intelligent Decision Support Systems
Chapter 4: Intelligent Decision Support Systems
4.1 Artificial Intelligence
4.1.1 AI Paradigms
4.1.1.1 Deliberative Approaches
4.1.1.1.0 Logic Paradigm
4.1.1.1.0 Heuristic Search Paradigm
4.1.1.1.0 Knowledge-Based Paradigm
4.1.1.1.0 Model-Based Paradigm
4.1.1.1.0 Experience-Based Paradigm
4.1.1.2 Reactive Approaches
4.1.1.2.0 Connectionism Paradigm
4.1.1.2.0 Evolutionary Computation Paradigm
4.1.1.2.0 Other Optimization Paradigms
4.1.1.3 Uncertainty Reasoning Models
4.1.1.3.0 Bayesian Networks
4.1.1.3.0 Fuzzy Logic Systems
4.2 IDSS Typology
4.3 Classification of IDSS
4.3.1 Model-Driven IDSSs
4.3.2 Data-Driven IDSSs
4.4 Conceptual Components of an IDSS
4.5 Considerations and Requirements of an IDSS
4.6 IDSS Architecture
4.7 IDSS Analysis, Design, and Development
4.8 IDSS Evaluation
4.9 Development of an IDSS: A First Example
References
Further Reading
Chapter 5: Model-Driven Intelligent Decision Support Systems
5.1 Introduction
5.2 Agent-Based Simulation Models
5.2.1 Multi-Agent Systems
5.2.1.1 The Belief-Desire-Intention Model
5.2.1.2 Agent Architectures
5.2.1.3 Communication
5.2.1.4 Cooperation
5.2.1.5 Coordination
5.2.1.6 Design and Development
5.2.1.7 Multi-agent Applications
5.2.2 Agent-Based Simulation
5.2.2.1 Deployment and Use
5.2.2.2 An Example of an Agent-Based Simulation Model
5.2.2.2.0 Analysis of the System
5.2.2.2.0 Identification of the Entities
5.2.2.2.0 Types of Agents
5.2.2.2.0 Identification of Scenarios/Strategies
5.2.2.2.0 Evaluation of the Simulation Results
5.3 Expert-Based Models
5.3.1 Fact Base
5.3.2 Knowledge Base
5.3.2.1 Modularization of the KB
5.3.3 Reasoning Component: The Inference Engine
5.3.3.1 Reasoning Cycle
5.3.3.1.0 Detection of Candidate Rules
5.3.3.1.0 Selection of the rule to be applied
5.3.3.1.0 Application of the Selected Rule
5.3.3.1.0 End of the Cycle
5.3.3.2 Inference Engines
5.3.3.2.0 Forward Reasoning
Example 1
Example 2
5.3.3.2.0 Backward Reasoning
Example 1
Example 2a and 2b
5.3.4 Meta-Reasoning Component
5.3.4.1 Reasoning Cycle with Meta-Rules
5.3.4.2 Hybrid Reasoning
5.3.4.2.0 Example
5.3.5 User Interface
5.3.6 Explanation Module
5.3.7 Knowledge Acquisition Module
5.3.8 Knowledge Engineer Interface
5.3.9 The Knowledge Engineering Process
5.3.10 An Example of an Expert-Based Model
5.3.10.1 Identification
5.3.10.2 Conceptualization
5.3.10.3 Formalization
5.3.10.4 Implementation
5.3.10.5 Testing
5.4 Model-Based Reasoning Methods
5.4.1 Introduction
5.4.2 Preliminaries
5.4.3 Consistency-Based Diagnosis
5.4.4 Abductive Diagnosis
5.4.5 Conclusions
5.5 Qualitative Reasoning Models
5.5.1 Basic Principles of Qualitative Reasoning
5.5.2 General Flowchart of a Qualitative Reasoning Model
5.5.3 Qualitative Model Building
5.5.3.1 Qualitative Model Formulation
5.5.3.1.0 Scenario
5.5.3.1.0 Ontologies
5.5.3.1.0 Component/Model Fragment Library
5.5.3.1.0 Model Formulation
5.5.3.1.0 Compositional Modelling
5.5.3.2 Qualitative Model Representation
5.5.3.2.0 Representing Continuous Magnitudes as Qualitative Values
Status Abstraction
Sign Algebra
Quantity Space
Qualitative Values
Interval Representation
Finite Algebras
5.5.3.2.0 Representing Mathematical Relationships
Confluences
Influences
Mathematical Functions
5.5.4 Qualitative Model Simulation
5.5.5 Main Qualitative Reasoning Frameworks
5.5.6 An Example of a Qualitative Reasoning Model
5.5.6.1 Qualitative Model Formulation
5.5.6.2 Qualitative Model Representation
5.5.6.2.0 Indirect Influences or Proportionalities
5.5.6.2.0 Direct Influences
5.5.6.2.0 Correspondences
5.5.6.2.0 Inequalities
5.5.6.3 Qualitative Model Simulation
References
Further Reading
Chapter 6: Data-Driven Intelligent Decision Support Systems
6.1 Introduction
6.2 Data Mining, Knowledge Discovery, and Data Science
6.2.1 Terminology in Data Mining
6.3 Pre-Processing Techniques
6.3.1 Data Fusion and Merge
6.3.2 Meta-Data Definition and Analysis
6.3.3 Data Filtering
6.3.4 Special Variables Management
6.3.4.1 Compositional Variables
6.3.4.2 Multi-valued Variables
6.3.5 Visualization and Descriptive Statistical Analysis
6.3.6 Transformation and Creation of Variables
6.3.6.1 Transformation of Existing Variables
6.3.6.2 Creation of New Variables
6.3.7 Outlier Detection and Management
6.3.8 Error Detection and Management
6.3.9 Missing Data Management
6.3.10 Data Reduction
6.3.10.1 Filters and Wrapper Methods
6.3.10.2 Instance Selection
6.3.10.3 Feature Selection
6.3.11 Feature Relevance
6.3.11.1 Relevance Detection
6.3.11.2 Feature Weighting
6.4 Data Mining Methods
6.4.1 Unsupervised Models
6.4.1.1 Descriptive Models
6.4.1.1.0 Partitional Clustering Techniques
K-Means Clustering
G-Means Clustering
Nearest-Neighbour Clustering
6.4.1.1.0 Hierarchical Clustering Techniques
Agglomerative/Ascendant Techniques
Divisive/Descendent Techniques
6.4.1.1.0 Validation of Descriptive Models
Structural Validation of Clusters
Qualitative Validation of Clusters
6.4.1.1.0 An Example of a Descriptive Model
6.4.1.2 Associative Models
6.4.1.2.0 Association Rules
Association Rule Methods
Validation of an Association Rule Model
An Example of an Association Rule Model
6.4.2 Supervised Models
6.4.2.1 Discriminant Models
6.4.2.1.0 Decision Trees
Information Gain Method
Gain Ratio Method
Impurity Measure Method
Tree Pruning
An Example of a Decision Tree Model
6.4.2.1.0 Case-Based Discriminant Models
Organization of Cases and the Case Library
Case Retrieval and Similarity Assessment
Case Adaptation
Case Evaluation
Case Learning
Case-Based Classifiers and Instance-Based Classifiers
An Example of a Case-Based Classifier
6.4.2.1.0 Ensemble Methods
Voting
Bagging
Random Forests
Boosting
An Example of an Ensemble Model
6.4.2.1.0 Validation of Discriminant Models
Estimation of the Error for Supervised Models
Quantitative Validation of Supervised Models
Validation Tools and Indicators in Discriminant Models
6.4.2.2 Predictive Models
6.4.2.2.0 Artificial Neural Network Models
The Perceptron, and the Basic Behaviour of an Artificial Neuron
Multi-Layer Perceptron and the Backpropagation Algorithm
An Example of an Artificial Neural Network Model
6.4.2.2.0 Case-Based Predictive Models
An Example of a Case-Based Predictive Model
6.4.2.2.0 Linear Regression Models
Multiple Linear Regression Models
Validation of a Multiple Linear Regression Model
An Example of a Multiple Linear Regression Model
6.4.2.2.0 An Ensemble of Predictive Models
Regression Trees and Predictive Random Forests
6.4.2.2.0 Validation of Predictive Models
6.4.3 Optimization Models
6.4.3.1 Genetic Algorithms
6.4.3.1.0 Chromosome Encoding
6.4.3.1.0 Fitness Function
6.4.3.1.0 Genetic Operators
6.4.3.1.0 General Scheme
6.5 Post-Processing Techniques
6.6 From Data Mining to Big Data
References
Further Reading
Chapter 7: The Use of Intelligent Models in Decision Support
7.1 Using Model-Driven Methods in IDSS
7.1.1 The Use of Agent-Based Simulation Models
7.1.2 The Use of Expert-Based Models
7.1.3 The Use of Model-Based Reasoning Techniques
7.1.4 The Use of Qualitative Reasoning Models
7.2 Using Data-Driven Methods in IDSS
7.2.1 The Use of Descriptive Models
7.2.2 The Use of Associative Models
7.2.3 The Use of Discriminant Models
7.2.4 The Use of Predictive Models
References
Further Reading
Part III: Development and Application of IDSS
Chapter 8: Tools for IDSS Development
8.1 Introduction
8.2 Tools for Data-Driven Methods
8.2.1 Weka
8.2.2 RapidMiner Studio
8.2.3 KNIME Analytics Platform
8.2.4 KEEL
8.2.5 Orange
8.2.6 IBM SPSS Modeler
8.2.7 SAS Enterprise Miner
8.3 Tools for Model-Driven Techniques
8.3.1 Agent-Based Simulation Tools
8.3.1.1 Prometheus
8.3.1.2 Jadex
8.3.1.3 Ascape
8.3.1.4 FLAME
8.3.1.5 Janus
8.3.1.6 Netlogo
8.3.1.7 Anylogic
8.3.2 Expert-Based Model Tools
8.3.2.1 CLIPS
8.3.2.2 Drools
8.3.2.3 Jess
8.3.3 Model-Based Reasoning Tools
8.3.3.1 Minion
8.3.3.2 Gecode
8.3.3.3 Choco Solver
8.3.4 Qualitative Reasoning Tools
8.3.4.1 Garp3
8.3.4.2 GQR
8.3.4.3 Simantics System Dynamics
8.4 General Development Environments
8.4.1 R
8.4.2 Python
8.4.3 GESCONDA
References
Further Reading
Chapter 9: Advanced IDSS Topics and Applications
9.1 Introduction
9.2 Uncertainty Management
9.2.1 Uncertainty Models
9.2.2 Pure Probabilistic Model
9.2.3 Certainty Factor Model
9.2.3.1 An Example Using the Certainty Factor Model
9.2.3.1.0 R1 Application
9.2.3.1.0 R2 Application
9.2.3.1.0 R3 Application
9.2.3.1.0 Combination (Co-conclusion) of the Two CFs
9.2.4 Bayesian Network Model
9.2.4.1 Fundamentals of Bayesian Networks
9.2.4.1.0 Dependence and Independence Relations in a Bayesian Network
9.2.4.1.0 Markov Condition and Factorization of the Joint Probability Distribution
9.2.4.1.0 Generalization of Conditional Independence Relations: D-Separation
9.2.4.2 Inference in Bayesian Networks
9.2.4.2.0 Exact Inference Methods
Marginalization or Summation Out Method
Enumeration Method
Variable Elimination Method
Evidence Propagation Through Local Message Passing Method
Inference in Multiply-Connected Networks
9.2.4.2.0 Approximate Inference Methods
Direct Sampling Methods
Markov Chain Monte Carlo Sampling Methods
Other Methods
9.2.5 Fuzzy Set Theory/Possibilistic Model
9.2.5.1 Fundamentals on Fuzzy Sets
9.2.5.2 Possibility Theory
9.2.5.3 Representation of Membership Functions and Linguistic Variables
9.2.5.4 Fuzzy Logic and Fuzzy Connectives
9.2.5.4.0 Fuzzy Connectives Operating on the Same Universe
9.2.5.4.0 Fuzzy Connectives Operating on Different Universes
9.2.5.5 Approximate Reasoning and Fuzzy Inference in a Rule-Based System
9.2.5.5.0 Fuzzy Inference
9.2.5.5.0 Fuzzy Inference with Precise Data: Fuzzy Control
The Mamdani Fuzzy Control Method
The Sugeno/Takagi-Sugeno Fuzzy Control Method
An Example Using the Mamdani Fuzzy Control Method
9.3 Temporal Reasoning Issues
9.3.1 The Temporal Reasoning Problem
9.3.2 Approaches to Temporal Reasoning
9.3.2.1 Dynamic Bayesian Networks
9.3.2.2 Temporal Artificial Neural Networks
9.3.2.3 Temporal Case-Based Reasoning
9.3.2.3.0 Episode-Based Reasoning (EBR)
9.3.2.3.0 Basic Terminology for EBR
9.3.2.3.0 Episode-Based Reasoning Memory Model
9.3.2.3.0 Episode Retrieval and Learning
9.3.2.4 Incremental Machine Learning Techniques and Data Stream Mining
9.4 Spatial Reasoning Issues
9.4.1 The Spatial Reasoning Problem
9.4.2 Approaches to Spatial Representation and Reasoning
9.4.3 Geographic Information Systems (GISs)
9.5 Recommender Systems
9.5.1 Formulation of the Problem
9.5.2 General Architecture of a Recommender System
9.5.3 Recommender System Techniques
9.5.3.1 Collaborative Filtering
9.5.3.2 Content-Based
9.5.3.3 Other Techniques
9.5.4 Evaluation of Recommender Systems
9.5.5 Applications of Recommender Systems
9.5.6 Future Trends in Recommender Systems
References
Further Reading
Chapter 10: Summary, Open Challenges, and Concluding Remarks
10.1 Summary
10.2 Open Challenges in IDSS
10.2.1 Integration and Interoperation of Models in IDSS
10.2.1.1 Interoperation and Model Interoperability
10.2.1.2 Tools and Techniques for Achieving Model Interoperability
10.2.2 An Interoperable Framework for IDSS Development
10.2.3 IDSS Evaluation
10.3 Concluding Remarks
References
Further Reading
Correction to: Intelligent Decision Support Systems
Index