Design of neural classi ers using variable
نویسندگان
چکیده
In this paper we present a new procedure for the design and optimization of a kind of artiicial neural networks (ANN) based on genetic algorithms (GAs). The method proposed here, called g-lvq , uses a genetic algorithm with variable-length genome and a vectorial tness to optimize Learning Vector Quantization (lvq) neural networks. The procedure optimizes simultaneously the classiication accuracy, the size of the networks, and the mean square error from the neural network to the training samples. After a description of the algorithm, results on synthetic and real-world problem are shown and discussed.
منابع مشابه
Improving Naive Bayes Classiiers Using Neuro-fuzzy Learning 1
Naive Bayes classi ers are a well-known and powerful type of classi ers that can easily be induced from a dataset of sample cases. However, the strong conditional independence and distribution assumptions underlying them can sometimes lead to poor classi cation performance. Another prominent type of classi ers are neuro-fuzzy classi cation systems, which derive (fuzzy) classi ers from data usin...
متن کاملA theory of classifier combination: the neural network approach
There is a trend in recent OCR development to improve system performance by combining recognition results of several complementary algorithms. This thesis examines the classi er combination problem under strict separation of the classi er and combinator design. None other than the fact that every classi er has the same input and output speci cation is assumed about the training, design or imple...
متن کاملUsing A Neural Network to Approximate An Ensemble of Classi ers
Several methods e g Bagging Boosting of constructing and combining an ensemble of classi ers have recently been shown capable of improving accuracy of a class of commonly used classi ers e g decision trees neural networks The ac curacy gain achieved however is at the expense of a higher requirement for storage and computation This storage and computation overhead can decrease the utility of the...
متن کاملComparison of Neural and Statistical Classiiers| Theory and Practice
Pattern classi cation using neural networks and statistical methods is discussed. We rst give a tutorial overview that groups popular classi ers according to their underlying mathematical principles into several distinct categories. Starting from the Bayes classi er, one division is whether the classi er is explicitly estimating class conditional densities, or directly estimating the posterior ...
متن کاملBias-Variance Analysis of ECOC and Bagging Using Neural Nets
One of the methods used to evaluate the performance of ensemble classi ers is bias and variance analysis. In this paper, we analyse bagging and ECOC ensembles using bias-variance domain of James [1] and make a comparison with single classi ers, when using Neural Networks (NNs) as base classi ers. As the performance of the ensembles depends on the individual base classi ers, it is important to u...
متن کامل