Showing papers by "Lakhmi C. Jain published in 1998"

Industrial Applications of Neural Networks

Abstract: From the Publisher: Industrial Applications of Neural Networks explores the success of neural networks in different areas of engineering endeavors. Each chapter shows how the power of neural networks can be exploited in modern engineering applications.

Cross-validation with active pattern selection for neural-network classifiers

Abstract: We propose a new approach for leave-one-out cross-validation of neural-network classifiers called "cross-validation with active pattern selection" (CV/APS). In CV/APS, the contribution of the training patterns to network learning is estimated and this information is used for active selection of CV patterns. On the tested examples, the computational cost of CV can be drastically reduced with only small or no errors.

Intelligent Adaptive Control: Industrial Applications

Abstract: From the Publisher: Intelligent Adaptive Control describes important techniques, developments, and applications of computational intelligence in system control. Intelligent Adaptive Control provides a state-of-the-art treatment of practical applications of computational intelligence in system control. The book cohesively covers introductory and advanced theory, design, implementation, and industrial useserving as a singular resource for the theory and application of intelligent control, particularly employing fuzzy logic, neural networks, and evolutionary computing.

Generation of fuzzy decision trees by fuzzy ID3 with adjusting mechanism of AND/OR operators

Abstract: An ability of adjusting the AND/OR operators is discussed in this paper to overcome the drawback of fuzzy ID3. The fuzzy ID3 has a mechanism to construct fuzzy rules. However, the fuzzy ID3 has a couple of problems, i.e., a lack of representation and an adjusting problem. In our fuzzy ID3, AND/OR operators are formulated using t-norm and t-conorm connectives with parameters and each parameter is adjusted using the golden section method. By using the golden section method, the proposed fuzzy ID3 gives accurate fuzzy rules. If the t-conorm connective is selected as AND/OR operator in the adjusting stage, the decision tree has more flexible representation.

Acquisition of fuzzy knowledge by NN and GA-A survey of the fusion and union methods proposed in Japan

Abstract: Fuzzy control is an effective method widely employed in the complex control field. However, in the fuzzy control, the adjustment (tuning) of the shape of membership functions and the acquisition of fuzzy rules is still a very active research field. To solve the tuning problem, a number of fusion/union methods using neural networks (NN) or genetic algorithms (GA) have been proposed. By scanning through these methods, we find that these methods are roughly divided into some categories depending on the grade of fusion/union between fuzzy models and NN, GA. In this papers, the fusion/union models are discussed and their applications in the field of control are explained.

On the effect of multi-parents recombination in binary coded genetic algorithms

Abstract: The recombination operator plays a very important role in genetic algorithms. We present binary coded genetic algorithms in which more than two parents are involved in the recombination operation. We propose two types of multi-parent recombination operators, the multi-cut (MX) and seed crossover (SX). Each of these operators is a natural generalization of two parents recombination operator. These operators are evaluated on the De Jong standard test functions. The results showed clearly that the multi-parent recombinations lead to better performance, although the performance improvement for different techniques were found to be dependent on problems.

Properties of robust solution searching in multi-dimensional space with genetic algorithms

Abstract: A large number of studies on genetic algorithms (GAs) emphasize finding a globally optimal solution. Some other investigations have also been made for detecting multiple solutions. If a global optimal solution is very sensitive to noise or perturbations in the environment then there may be cases where it is not good to use this solution. We have proposed a new scheme, GA/RS/sup 3/, which extends the application of GAs to domains that require the discovery of robust solutions and a mathematical model for this scheme has been developed restricting their search space to one-dimensional. We analyze properties of GAs/RS/sup 3/ in multi-dimension search spaces. The effectiveness of the scheme is demonstrated by solving two-dimensional functions having broad and sharp peaks.

Degree of familiarity ART2 in fuzzy fusion landmine detection

Abstract: The self organising network ART2 has been modified to provide a fuzzy output value, which indicated the degree of familiarity of a new analogue input pattern to previous patterns stored in the long term memory weights of the network The outputs of the multilayer perceptron and this modification output of ART2 allows it to now provide an analogue value to a fuzzy rule-based fusion technique which also uses a processed polarisation resolved image as its third input In real-time these two classifier outputs indicate the likelihood of a surface landmine target when presented with a number of multispectral and textural bands Due to the modifications in ART2 this updated alternative architecture has improved real-time landmine detection capabilities although the registration of all bands is more critical to the accuracy of results in this case The real-time fuzzy rule based system has detected two of the 3 landmines and the landmine surrogate with 2 false alarms