Skip to main content
main-content

Über dieses Buch

Develops insights into solving complex problems in engineering, biomedical sciences, social science and economics based on artificial intelligence. Some of the problems studied are in interstate conflict, credit scoring, breast cancer diagnosis, condition monitoring, wine testing, image processing and optical character recognition. The author discusses and applies the concept of flexibly-bounded rationality which prescribes that the bounds in Nobel Laureate Herbert Simon’s bounded rationality theory are flexible due to advanced signal processing techniques, Moore’s Law and artificial intelligence.

Artificial Intelligence Techniques for Rational Decision Making examines and defines the concepts of causal and correlation machines and applies the transmission theory of causality as a defining factor that distinguishes causality from correlation. It develops the theory of rational counterfactuals which are defined as counterfactuals that are intended to maximize the attainment of a particular goal within the context of a bounded rational decision making process. Furthermore, it studies four methods for dealing with irrelevant information in decision making:

Theory of the marginalization of irrelevant information Principal component analysis Independent component analysisAutomatic relevance determination method

In addition it studies the concept of group decision making and various ways of effecting group decision making within the context of artificial intelligence.

Rich in methods of artificial intelligence including rough sets, neural networks, support vector machines, genetic algorithms, particle swarm optimization, simulated annealing, incremental learning and fuzzy networks, this book will be welcomed by researchers and students working in these areas.

Inhaltsverzeichnis

Frontmatter

1. Introduction to Rational Decision Making

Abstract
This chapter studies causality, correlation and artificial intelligence for decision making. Correlation methods are defined and applied to build correlation machines which are used for rational decision making. Furthermore, causality concepts are defined and used to build causal machines which are also used for rational decision making. Artificial intelligence methods considered in this book are support vector machines, rough sets, fuzzy systems, neural networks, particle swarm optimization, the genetic algorithm and simulated annealing.
Tshilidzi Marwala

2. Causal Function for Rational Decision Making: Application to Militarized Interstate Dispute

Abstract
This chapter describes and defines a causal function within the context of rational decision making. This is implemented using rough sets to build the causal machines. The rough sets were successfully used to identify the causal relationship between the militarized interstate dispute variables (causes) and conflict status effects.
Tshilidzi Marwala

3. Correlation Function for Rational Decision Making: Application to Epileptic Activity

Abstract
This chapter describes and defines a correlation function within the context of rational decision making. The correlation function is implemented using support vector machine. It was successfully applied to identify the correlation relationship between the electroencephalogram (EEG) signal and eliptic activity of patients.
Tshilidzi Marwala

4. Missing Data Approaches for Rational Decision Making: Application to Antenatal Data

Abstract
This chapter introduces missing data estimation for rational decision making. In this chapter it is assumed that there is a fixed topological characteristic between the variables required to make a rational decision and the actual rational decision. This, therefore, implies that rational decision making can be viewed as a missing data in a topology that includes both the action variables and the decision. This technique is applied using an autoassociative multi-layer perceptron network trained using scaled conjugate method and the missing data is estimated using genetic algorithm. This technique is used to predict HIV status of a subject given the demographic characteristics.
Tshilidzi Marwala

5. Rational Counterfactuals and Decision Making: Application to Interstate Conflict

Abstract
This chapter introduces the concept of rational counterfactuals which is an idea of identifying a counterfactual from the factual (whether perceived or real), and knowledge of the laws that govern the relationships between the antecedent and the consequent, that maximizes the attainment of the desired consequent. In counterfactual thinking factual statements like: ‘Saddam Hussein invaded Kuwait and consequently George Bush declared war on Iraq’ and with its counterfactual being: ‘If Saddam Hussein did not invade Kuwait then George Bush would not have declared war on Iraq’. In this chapter in order to build rational counterfactuals neuro-fuzzy model and genetic algorithm are applied. The theory of rational counterfactuals is applied to identify the antecedent that gives the desired consequent necessary for rational decision making. The rational counterfactual theory is applied to identify the values of variables Allies, Contingency, Distance, Major Power, Capability, Democracy, as well as Economic Interdependency that give the desired consequent Peace.
Tshilidzi Marwala

6. Flexibly-Bounded Rationality in Interstate Conflict

Abstract
This chapter applies the theory of flexibly bounded rationality to interstate conflict. Flexibly bounded rationality is a theory that states that the bounds prescribed by Herbert Simon in his theory of bounded rationality are flexible. On contextualizing the theory of flexibly bounded rationality, inference, the theory of rational expectation, the theory of rational choice and the theory of rational conterfactuals are described. The theory of flexibly bounded rationality is applied for decision making process. This is done by using a multi-layer perceptron network and particle swarm optimization.
Tshilidzi Marwala

7. Filtering Irrelevant Information for Rational Decision Making

Abstract
This chapter deals with the concept of using relevant information as a basis of rational decision making. In this regard, whenever information is irrelevant it needs to be marginalized or eliminated. Making decisions using information which contains irrelevant information often confuses a decision making process. In this chapter we discuss four methods for making rational decisions by either marginalizing irrelevant information or not using irrelevant information. These methods are marginalization of irrationality approach, automatic relevance determination, principal component analysis and independent component analysis. These techniques are applied to condition monitoring, credit scoring, interstate conflict and face recognition.
Tshilidzi Marwala

8. Group Decision Making

Abstract
Decision making by individuals and its relation to rationality is an important area of research because it allows individuals to make effective decisions. Decision making in groups is when individuals come together and make decisions in whatever manner that is feasible. Governments, organizations and companies are involved in group decision making. This chapter studies the concept of group decision making and how artificial intelligence is used to facilitate decision making in a group. Four group based decision making techniques are considered and these are ensemble of support vector machines which are applied to land cover mapping, incremental learning using genetic algorithm which is applied to optical character recognition, dynamically weighted mixtures of experts which are applied to platinum price prediction as well as Learn++ which is applied to wine recognition.
Tshilidzi Marwala

9. Conclusion

Abstract
This book was on rational decision making with the aide of artificial intelligence. The classical definition of a rational agent is an agent which acts to maximize its utility. Utility is a difficult concept to grasp. It is classically defined as the ability of an object to satisfy needs. Much has been written about utility and its derivative expected utility. One definition of utility which dominates the economics field is that it is a representation of the preference of some goods or services. Samuelson attempted to quantify utility as a measure of the willingness of the people to pay for a particular good. In this book we view utility as a measure of the value of some goods less the cost associated with the acquisition of such goods. This simply means that we define utility as the value that is derived from a good minus the cost of that good. Thus if the good is quite valuable but is also equally expensive then its utility is zero because its value and cost balance out.
Tshilidzi Marwala

Backmatter

Weitere Informationen

Premium Partner

    Bildnachweise