Intelligent Systems and Applications in Business and Finance (Studies in Fuzziness and Soft Computing Book 415)

This book presents a selection of current research results in the field of intelligent systems and draws attention to their practical applications and issues connected with the areas of decision-making, economics, business and finance. The nature of the contributions is interdisciplinary – combining psychological and behavioural aspects with the theory and practice of decision-support, design of intelligent systems and development of machine learning tools. The authors, among other topics, discuss the multi-expert evaluation with intangible criteria, suggest a redefinition of the standard multiple-criteria decision-making framework, propose novel methods for causal map analysis and new feature selection methods. The topics are selected to stress the potential of the up-to-date intelligent methods to deal with practical problems relevant in these areas and to provide inspiration for advanced students, researchers and practitioners in the respective fields. Preface Contents Attitude-Based Multi-expert Evaluation of Design 1 Introduction 2 Semantic Differential 3 Interval-Valued Semantic Differential 4 Evaluation Using the Interval-Valued Semantic Differential, Consensus of Evaluations 5 Conclusion References An Investigation of Hidden Shared Linkages Among Perceived Causal Relationships in Cognitive Maps 1 Introduction 2 Data, Theoretical Aspects, and Methodology 2.1 Data Sample 2.2 Cognitive Maps 2.3 Inter-causal Relationships 2.4 Set-Theoretic Consistency and Coverage Measures 2.5 Hypothesis Models 2.6 Research Process 3 Results and Discussion 3.1 Inter-causal Relationships Under H1 3.2 Inter-causal Relationships Under H2 4 Conclusion and Future Directions References A New Framework for Multiple-Criteria Decision Making: The Baseline Approach 1 Introduction 2 Definition of a Standard MCDM Problem 3 Definition of the New MCDM Problem Addressed in This Paper—Decision Making Under the Existence of a Baseline 3.1 The Basic Difference Between the Standard Formulation of the MCDM Problem and the Newly Proposed Generalized Formulation 3.2 Assumptions and Notation for the Newly Proposed MCDM Problem Formulation 4 Decision-Making Styles Available with the New MCDM Problem Formulation 4.1 ``Appreciating Possessed'' Decision-Making Style 4.2 ``Craving Unavailable'' Decision-Making Style 4.3 A Combined Style—``Appreciating Possessed'' in the Baseline and ``Craving Unavailable'' in Alternatives 5 Behavioral Effects in the New MCDM Problem Formulation When the Combined Style of Criteria Weights Determination is Used 5.1 Cost of Switching Effect 5.2 (Un)Satisfaction Effect 5.3 Constant Switching Effect 5.4 New Feature Effect 6 Conclusions References Fuzzy Similarity and Entropy (FSAE) Feature Selection Revisited by Using Intra-class Entropy and a Normalized Scaling Factor 1 Introduction 2 Methods 2.1 Class-Wise Fuzzy Similarity and Entropy (C-FSAE) Feature Selection 2.2 Class-Wise Fuzzy Entropy and Similarity (C-FES) Feature Selection 3 Data 3.1 Artificial Example Data Sets 3.2 Real-World Data Sets 4 Training Procedure 5 Results 6 Conclusion References A Region-Based Approach for Missing Value Imputation of Cooling Technologies for the Global Thermal Power Plant Fleet Using a Decision Tree Classifier 1 Introduction 2 Data and Methodology 2.1 Power Plant Data 2.2 Decision Tree Classifier 2.3 Different Approaches to Account for the Geographical Location of Power Plants 2.4 Overcoming the Problem of Data Scarcity 2.5 Validation of Results 3 Results 3.1 Country Model Versus Minor Region Model 3.2 Performance of Hybrid Model 4 Conclusion References A Neural Network Based Multi-class Trading Strategy for the S&P 500 Index 1 Introduction 2 Data 2.1 Data Gathering and Feature Engineering 2.2 Missing Values 2.3 Target Variables 3 Feature Selection 3.1 Information Gain 3.2 Pearson Correlation 3.3 Random Forest Feature Importance 3.4 Feature Importance Results 4 Model Performance Evaluation 4.1 Artificial Neural Networks 4.2 Model Evaluation 4.3 Model Selection 5 Results 5.1 Trading Strategies and Threshold 5.2 Performance 6 Conclusion Appendix References Predicting Short-Term Traffic Speed and Speed Drops in the Urban Area of a Medium-Sized European City—A Traffic Control and Decision Support Perspective 1 Introduction 2 A Brief Overview of Previous Studies 3 Data Set and Applied Methods 3.1 Speed Prediction in Helsinki 3.2 Models Used in the Analysis 4 Results for Traffic Speed Prediction and Their Discussion 4.1 ARMA Models for Traffic Speed Prediction 4.2 Linear Regression Models for Traffic Speed Prediction 4.3 K-Nearest Neighbors Method for the Prediction of Traffic Speed 4.4 XGBoost for the Prediction of Traffic Speed 4.5 Comparison of the Performance of ARMA, Linear Regression, KNN and XGBoost Models for Traffic Speed Prediction 5 Speed Drop Prediction 5.1 Speed Drop Prediction Capabilities by the Analysed Models 5.2 A Decision Tree Based Prediction of Traffic Jams 6 Conclusions References Hedging Effectiveness of Currency ETFs Against WTI Crude Oil Price Fluctuations 1 Introduction 2 Review of the Literature 3 Methodology 3.1 ARMA-GARCH Modelling 3.2 The Copulas 3.3 Hypotheses and Testing 4 Results 4.1 Descriptive Statistics 4.2 Margins Modeling with EGARCH 4.3 Copula Modeling Results 5 Conclusion References Chapter 1: Attitude-based multi-expert evaluation of design -- Chapter 2: An investigation of hidden shared linkages among perceived causal relationships in cognitive maps -- Chapter 3: A new framework for multiple-criteria decision making: The baseline approach -- Chapter 4: Fuzzy similarity and entropy (FSAE) feature selection revisited by using intra-class entropy and a normalized scaling factor -- Chapter 5: A region-based approach for missing value imputation of cooling technologies for the global thermal power plant fleet using a decision tree classifier -- Chapter 6: A neural network based multi-class trading strategy for the S&P 500 index -- Chapter 7: Predicting short-term traffic speed and speed drops in the urban area of a medium-sized European city - a traffic control and decision support perspective -- Chapter 8: Hedging effectiveness of currency ETFs against WTI crude oil price fluctuations