Data Mining with Computational Intelligence (Advanced by Lipo Wang

By Lipo Wang

Discovering details hidden in info is as theoretically tough because it is essentially vital. With the target of learning unknown styles from facts, the methodologies of information mining have been derived from statistics, desktop studying, and synthetic intelligence, and are getting used effectively in program components corresponding to bioinformatics, banking, retail, and so forth. Wang and Fu found in element the state-of-the-art on find out how to make the most of fuzzy neural networks, multilayer perceptron neural networks, radial foundation functionality neural networks, genetic algorithms, and aid vector machines in such functions. They specialize in 3 major facts mining projects: facts dimensionality aid, category, and rule extraction. The booklet is concentrated at researchers in either academia and undefined, whereas graduate scholars and builders of information mining platforms also will cash in on the specific algorithmic descriptions.

Show description

Read Online or Download Data Mining with Computational Intelligence (Advanced Information and Knowledge Processing), 1st Edition PDF

Similar game theory books

The Evolution of Cooperation (Revised Edition)

Contributor be aware: ahead by way of Richard Dawkins

The Evolution of Cooperation presents important insights into the age-old query of even if unforced cooperation is ever attainable. commonly praised and much-discussed, this vintage ebook explores how cooperation can emerge in an international of self-seeking egoists-whether superpowers, companies, or individuals-when there isn't any valuable authority to police their activities.

The challenge of cooperation is relevant to many alternative fields. Robert Axelrod recounts the recognized desktop tournaments during which the “cooperative” software Tit for Tat recorded its lovely victories, explains its software to a wide spectrum of topics, and indicates how readers can either observe cooperative rules to their very own lives and train cooperative ideas to others.

The Mathematics of Arbitrage (Springer Finance)

Evidence of the "Fundamental Theorem of Asset Pricing" in its common shape via Delbaen and Schachermayer was once a milestone within the heritage of contemporary mathematical finance and now varieties the cornerstone of this ebook. places into booklet structure a chain of significant effects due more often than not to the authors of this publication. Embeds highest-level study effects right into a therapy amenable to graduate scholars, with introductory, explanatory historical past.

Multicriteria Analysis: Applications to Water and Environment Management

Multicriteria research is likely one of the most vital fields of selection technology. This e-book offers an summary of the formula of an acceptable version and provides a finished precis of the preferred equipment for fixing multicriteria selection difficulties. as well as the classical procedure the booklet introduces fuzzy and stochastic method, types with uncertainty, social selection and clash answer.

Extra resources for Data Mining with Computational Intelligence (Advanced Information and Knowledge Processing), 1st Edition

Sample text

The simulations indicate that the input points 1, 4, and 5 are consistently less important than other inputs (Fig. 5). Simulations are re-run after these less important inputs are eliminated. This results in a network of size 17:2:1 (seventeen inputs, two hidden neurons and one output neuron). 1 Wavelet MLP Neural Networks for Time-series Prediction 31 Fig. 5. Distribution of relative importance of 20 inputs for the wavelet MLP network with decomposition level one in one of the simulations, which is similar to the results in other simulations.

An introduction to data mining methods is presented. In Chap. 2, a wavelet multi-layer perceptron neural network is described for predicting temporal sequences. The multi-layer perceptron neural network has its input signal decomposed to various resolutions using a wavelet transformation. The time frequency information which is normally hidden is exposed by the wavelet transformation. Based on the wavelet transformation, less important wavelets are eliminated. Compared with the conventional MLP network, the wavelet MLP neural network has less performance swing sensitivity to weight initialization.

After the less important inputs are eliminated, the wavelet MLP shows more consistent performance for diļ¬€erent weight initializations in comparison to the conventional MLP [303][332]. 1 Introduction to Wavelet Multi-layer Neural Network A time-series is a sequence of data that vary with time, for example, the daily average temperature from the year 1995 to 2005. The task of time-series prediction is to forecast future trend using the past values in the time-series. There exist many approaches to time-series prediction.

Download PDF sample

Rated 4.78 of 5 – based on 36 votes