Computational intelligence

This book is an introduction to computational intelligence. Its content follows the curriculum of Computational Intelligence in Economics for the 3rd year undergraduate students, specialization Economic Cybernetics.

Chapter 1 contains definitions and results on Zadeh’s fuzzy sets theory: fuzzy sets, fuzzy relations, their operations and their application to multicriteria decision making problem are presented. The second chapter is dedicated to applying fuzzy numbers of economic modeling. After the presentation of fuzzy numbers and associated possibilistic indicators, in the chapter several possibilistic or mixed models of investment risk are developed.

A very brief chapter describes data exploration in R. Chapter 4 deals with principal component analysis. In Chapter 5 non-hierarchical clustering methods are presented: (k-means algorithm, fuzzy k-means algorithm, k-medoids algorithm) and hierarchical clustering method (agglomerative hierarchical algorithm).

Chapter 6 contains a few of the well-known classification algorithms: Naïve Bayesian Classifier, the K-nearest neighbor algorithm, SVM, decision trees. A section of the chapter is dedicated to some performance indicators of classifiers and their comparison. Parameter estimation by EM (expectation maximization) is discussed in Chapter 7.

Chapter 8 is a very brief introduction to neural networks. After the definitions of artificial neuron, artificial neural networks and the learning process, some of the known learning rules are surveyed: the perceptron rule, Hebb’s rule, Oja and Sanger learning rules, the delta rule and the generalized delta rule (back-propagation). A section of the chapter discusses Kohonen neural networks.

Chapter 9 contains a summary description of the structure and components of a genetic algorithm: genetic codification, population initialization, selection operations, crossover and mutation, etc.

The models and algorithms are accompanied by numerical examples and applications such as R codes.