No label defined (Q4955558)

In the computer epoch, storming data are obtained every second. Informations need to be mined automatically at high speed. While computational techniques appear diversely in different fields. Coming up with new computational techniques, the collective techniques of different computational approaches have appeared as data fusion, data mining and so on. One expects a book concerning the advances in relational and hybrid methods for data mining both in algorithms and in mathematics. The present book meets the former demands. Though the involved approaches focus on financial applications, they are available in other fields as well. However, the present one is a heavily ``references pasted'' survey book. In this book, it usually shows you how to use a specified algorithm step by step and does not tell you why. If the reader wants to know more about the algorithms, what is the mathematical background and formula, why it works, how good it is, then he or she is refered to the references. Yet we can not expect a huge volume book including all aspects we want to know. In this book, as parts of the ``collective'' data mining, ARIMA, neural networks, decision tree, Markov models, hybrid knowledge-based networks and hybrid relational methods are examined for modelling and forecasting financial time series, guidance for new predictive and decision-enhancing hybrid methods to capital-market investments, trading, banking services and so on is then provided. Now let us take a close look at its contents: Chapter 1 -- The Scope and Methods of the Study. General ideas of data mining are presented. Chapter 2 -- Numerical Data Mining Models with Financial Applications. ARIMA, Instance-based learning, recurrent networks, hybrid systems, genetic algoritms testing methods, expert mining, interractive learning and so on are illustrated. Chapter 3 -- Rule based and Hybrid Financial Data Mining. Decision tree extracted from neural networks, disjunctive normal form learning, stochastic models, mixture of experts are discussed. Chapter 4 -- Relational Data Minig (RDM). A mechine method for discovering regularities and its extentions are emphasized. Chapter 5 -- Financial Applications of RDM. Chapter 6 -- Comparison of Performance of RDM and other methods in financial applications. Chapter 7 -- Fuzzy logic approach and its financial applications. Nevertheless, this book will be more readable if some shortcomings are avoided. These are, e.g. Too many abbreviations which come at the first appearance site. They are even not explained in the Subject Index. Some cited papers do not appear in the References; for example on page 110, the paper of Demper et al. (1976) which is an important reference for the EM algorithm) does not.