PubMed Health⌕ Search

SEARCH · PubMed Health

Results for “Multilayer Perceptrons”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 55 records · Page 3Linked to original sources
PubMed2006

Prediction for energy content of Taiwan municipal solid waste using multilayer perceptron neural networks.

In the past decade, the treatment amount of municipal solid waste (MSW) by incineration has increased significantly in Taiwan. By year 2008, approximately 70% of the total MSW generated will be incinerated. The energy content (usually expressed by lower heating value [LHV]) of MSW is an important parameter for the selection of incinerator capacity. In this work, wastes from 55 sampling sites, including villages, towns, cities, and remote islands in the Taiwan area, were sampled and analyzed once a season from April 2002 to March 2003 to determine the waste characteristics. The LHV of MSW in Taiwan was predicted by the multilayer perceptron (MLP) neural networks model using the input parameters of elemental analysis and dry- or wet-base physical compositions. Although all three of the models predicted LHV values rather accurately, the elemental analysis model provided the most accurate prediction of LHV values. Additionally, the wet-base physical composition model was the easiest and most economical. Therefore, the waste treatment operators can choose the more appropriate analysis method considering situations themselves, such as time, equipment, technology, and cost.

Computer Simulation↗
PubMed2005

Convective oxygen transport in a constructed wetland pond: mechanism, measurements and modelling by multilayer perceptrons.

The study presented here found evidence for the presence and importance of convective flows and associated oxygen transport in a constructed wetland pond in southern Finland. These flows are triggered by nightly cooling of the water at the surface, which may then become denser than the water lower down in the pond. The resulting layering (heavier water overlying less dense one) is hydromechanically unstable and--almost immediately--starts driving convective motion. This flow takes oxygen enriched water from the surface to the bottom of the wetland, where a distinct rise in oxygen saturation is recorded after some time lag. The process described can be modelled successfully by means of so-called Multilayer Perceptrons (MLPs), a class of Artificial Neural Networks. As explored in this study, these models are well suited to "learn" the mechanism of convective transport, which results in their ability to forecast oxygen saturation near the wetland bottom at a satisfactory level of accuracy.

Convection↗
PubMed1997

Computer image analysis of ultrasound images for discriminating and grading liver parenchyma disease employing a hierarchical decision tree scheme and the multilayer perceptron neural network classifier.

Differential diagnosis of liver parenchyma disease and grading of the hepatic disease on ultrasound is a common radiological problem that influences patient management. The aim of this study was to apply image analysis methods on ultrasound images for discriminating liver cirrhosis from fatty liver infiltration and for grading hepatic disease, which is important in the management of the patients. Ultrasound images of histologically confirmed 18 livers with cirrhosis, 37 livers with fatty infiltration, and 24 normal livers of healthy volunteers were selected and were digitized for further computer processing. Twenty two textural features were calculated from small matrix samples selected from the ultrasound image matrix of the liver parenchyma. These features were used in the design a three level hierarchical decision tree classification scheme, employing the multilayer perceptron neural network classifier at each hierarchical tree level. At the first tree level, classification accuracy for distinguishing normal from abnormal livers was 93.7%, at the second level the accuracy for discriminating cirrhosis from fatty infiltration was 90.9%, and at the third level the accuracy in distinguishing between low and high grading liver cirrhosis or fatty infiltration was 94.1% and 84.9% respectively. The proposed computer software system may be of value to the radiologists in assessing liver parenchyma disease.

Decision Trees↗
PubMed2004

Solid-phase microextraction, gas chromatography, and mass spectrometry coupled with discriminant factor analysis and multilayer perceptron neural network for detection of Escherichia coli.

This study was performed to investigate the ability of using discriminant factor analysis (DFA) and an artificial neural network (ANN) to identify and quantify the number of Escherichia coli (ATCC 25922) in nutrient media from data generated by analysis of E. coli volatile metabolic compounds using solid-phase microextraction (SPME) coupled with gas chromatography (GC) and mass spectrometry (MS). E. coli was grown in super broth and incubated at 37 degrees C for 2 to 12 h. Numbers of E. coli were followed using a colony counting method. An SPME device was used to collect the volatiles from the headspace above the samples, and the volatiles were identified using GC-MS. DFA was used to classify the samples from different incubation times. From DFA, it was possible to differentiate super broth from media containing E. coli when cell numbers were 10(5) CFU or more. The potential to predict the number of E. coli from the SPME-GC-MS data was investigated using a multilayer perceptron (MLP) neural network with back propagation training. The MLP comprised an input layer, one hidden layer, and an output layer, with a hyperbolic tangent sigmoidal transfer function in the hidden layer and a linear transfer function in the output layer. Good prediction was found as measured by a regression coefficient (R2 = 0.996) between actual and predicted data.

Colony Count, Microbial↗
PubMed1996

A Regularity Condition of the Information Matrix of a Multilayer Perceptron Network.

The Fisher information matrix of a multi-layer perceptron network can be singular at certain parameters, and in such cases many statistical techniques based on asymptotic theory cannot be applied properly. In this paper, we prove rigorously that the Fisher information matrix of a three-layer perceptron network is positive definite if and only if the network is irreducible; that is, if there is no hidden unit that makes no contribution to the output and there is no pair of hidden units that could be collapsed to a single unit without altering the input-output map. This implies that a network that has a singular Fisher information matrix can be reduced to a network with a positive definite Fisher information matrix by eliminating redundant hidden units. Copyright 1996 Elsevier Science Ltd

Journal Article↗
PubMed2000

Local minima and plateaus in hierarchical structures of multilayer perceptrons.

Local minima and plateaus pose a serious problem in learning of neural networks. We investigate the hierarchical geometric structure of the parameter space of three-layer perceptrons in order to show the existence of local minima and plateaus. It is proved that a critical point of the model with H - 1 hidden units always gives many critical points of the model with H hidden units. These critical points consist of many lines in the parameter space, which can cause plateaus in learning of neural networks. Based on this result, we prove that a point in the critical lines corresponding to the global minimum of the smaller model can be a local minimum or a saddle point of the larger model. We give a necessary and sufficient condition for this, and show that this kind of local minima exist as a line segment if any. The results are universal in the sense that they do not require special properties of the target, loss functions and activation functions, but only use the hierarchical structure of the model.

Models, Neurological↗
PubMed1997

Approximations of Functions by a Multilayer Perceptron: a New Approach.

We provide a radically elementary proof of the universal approximation property of the one-hidden layer perceptron based on the Taylor expansion and the Vandermonde determinant. It works for both L(q) and uniform approximation on compact sets. This approach naturally yields some bounds for the design of the hidden layer and convergence results (including some rates) for the derivatives. A partial answer to Hornik's conjecture on the universality of the bias is proposed. An extension to vector valued functions is also carried out. Copyright 1997 Elsevier Science Ltd.

Journal Article↗
PubMed2002

Systematic learning of gene functional classes from DNA array expression data by using multilayer perceptrons.

Recent advances in microarray technology have opened new ways for functional annotation of previously uncharacterised genes on a genomic scale. This has been demonstrated by unsupervised clustering of co-expressed genes and, more importantly, by supervised learning algorithms. Using prior knowledge, these algorithms can assign functional annotations based on more complex expression signatures found in existing functional classes. Previously, support vector machines (SVMs) and other machine-learning methods have been applied to a limited number of functional classes for this purpose. Here we present, for the first time, the comprehensive application of supervised neural networks (SNNs) for functional annotation. Our study is novel in that we report systematic results for ~100 classes in the Munich Information Center for Protein Sequences (MIPS) functional catalog. We found that only ~10% of these are learnable (based on the rate of false negatives). A closer analysis reveals that false positives (and negatives) in a machine-learning context are not necessarily "false" in a biological sense. We show that the high degree of interconnections among functional classes confounds the signatures that ought to be learned for a unique class. We term this the "Borges effect" and introduce two new numerical indices for its quantification. Our analysis indicates that classification systems with a lower Borges effect are better suitable for machine learning. Furthermore, we introduce a learning procedure for combining false positives with the original class. We show that in a few iterations this process converges to a gene set that is learnable with considerably low rates of false positives and negatives and contains genes that are biologically related to the original class, allowing for a coarse reconstruction of the interactions between associated biological pathways. We exemplify this methodology using the well-studied tricarboxylic acid cycle.

Algorithms↗
PubMed1993

Implantable cardioverter defibrillator electrogram recognition with a multilayer perceptron.

With the continuing development of new pacing and shock modalities, implantable cardioverter defibrillators must be able to recognize an increasingly wider range of arrhythmias so that therapy is delivered in an optimal way. Current rate-based systems can no longer meet this need. It is, therefore, necessary to develop new techniques that consider waveform morphology in addition to heart rate. Any developments, however, must be compatible with the strict power and space constraints imposed by implantable devices. Artificial neural networks offer a potentially viable way of meeting these objectives. In this article an artificial neural network based approach to the classification of arrhythmias from the ventricular intracardiac electrogram is described. It performs its classification on a set of easily extracted features that characterize waveform morphology and heart rate. Simulations showed that the artificial neural network performed better than a rate-based scheme similar to those used in some commercial devices.

Algorithms↗
PubMed2005

Prediction of caspase cleavage sites using Bayesian bio-basis function neural networks.

MOTIVATION: Apoptosis has drawn the attention of researchers because of its importance in treating some diseases through finding a proper way to block or slow down the apoptosis process. Having understood that caspase cleavage is the key to apoptosis, we find novel methods or algorithms are essential for studying the specificity of caspase cleavage activity and this helps the effective drug design. As bio-basis function neural networks have proven to outperform some conventional neural learning algorithms, there is a motivation, in this study, to investigate the application of bio-basis function neural networks for the prediction of caspase cleavage sites. RESULTS: Thirteen protein sequences with experimentally determined caspase cleavage sites were downloaded from NCBI. Bayesian bio-basis function neural networks are investigated and the comparisons with single-layer perceptrons, multilayer perceptrons, the original bio-basis function neural networks and support vector machines are given. The impact of the sliding window size used to generate sub-sequences for modelling on prediction accuracy is studied. The results show that the Bayesian bio-basis function neural network with two Gaussian distributions for model parameters (weights) performed the best and the highest prediction accuracy is 97.15 +/- 1.13%. AVAILABILITY: The package of Bayesian bio-basis function neural network can be obtained by request to the author.

Algorithms↗
PubMed1998

A neural network approach to the rapid computation of rotational correlation times from slow motional ESR spectra.

We explore the use of feed forward artificial neural networks for determining rotational correlation times from slow motional nitroxide electron spin resonance spectra. This approach is rapid and potentially eliminates the need for traditional iterative fitting procedures. Two networks are examined: the radial basis network and the multilayer perceptron. Although the radial basis network trains rapidly and performs well on simulated spectra, it is less satisfactory when applied to experimental spectra. In contrast, the multilayer perceptron trains slowly but is excellent at extracting correlation times from experimental spectra. In addition, the multilayer perceptron operates well in the presence of noise as long as the signal-to-noise ratio is greater than approximately 200/1. These findings suggest neural networks offer a promising approach for rapidly extracting correlation times without the need for iterative simulations.

Electron Spin Resonance Spectroscopy↗
PubMed1996

How dependencies between successive examples affect on-line learning.

We study the dynamics of on-line learning for a large class of neural networks and learning rules, including backpropagation for multilayer perceptrons. In this paper, we focus on the case where successive examples are dependent, and we analyze how these dependencies affect the learning process. We define the representation error and the prediction error. The representation error measures how well the environment is represented by the network after learning. The prediction error is the average error that a continually learning network makes on the next example. In the neighborhood of a local minimum of the error surface, we calculate these errors. We find that the more predictable the example presentation, the higher the representation error, i.e., the less accurate the asymptotic representation of the whole environment. Furthermore we study the learning process in the presence of a plateau. Plateaus are flat spots on the error surface, which can severely slow down the learning process. In particular, they are notorious in applications with multilayer perceptrons. Our results, which are confirmed by simulations of a multilayer perceptron learning a chaotic time series using backpropagation, explain how dependencies between examples can help the learning process to escape from a plateau.

Learning↗
PubMed2006

Use of artificial neural networks in near-infrared reflectance spectroscopy calibrations for predicting the inclusion percentages of wheat and sunflower meal in compound feedingstuffs.

The use of near-infrared reflectance spectroscopy (NIRS) calibrations to predict the ingredient composition in compound feeds (i.e., inclusion percentage of each ingredient) is a complex task, regarding both the nature of the parameters to be predicted, since they are not well-defined chemical entities, and the heterogeneousness of the matrices/formulas in which each ingredient participates. The present paper evaluates the use of nonlinear regression methods, such as artificial neural networks (ANN), for developing NIRS calibrations to predict these parameters. Two of the most representative ingredients in the Spanish compound feed formulations (wheat and sunflower meal) were selected for evaluating ANN possibilities, using a large spectral library comprising a total of 7523 commercial compound feed samples; 7423 were used as training set and 100 as validation set. Three general models of networks were studied: multilayer perceptron with back-propagation training (BP), multilayer perceptron with Levenberg-Maquartd training (LM), and radial basis function nets (RBF); moreover, in accordance with a factorial design, more complex architectures were evaluated gradually, changing the number of hidden layers and hidden neurons, for the determination of the optimal network topology. For both ingredients, the best results were obtained using ANN with BP training, showing prediction error values (SEP) of 2.72% and 0.66% for wheat and sunflower meal, respectively. These SEP values showed a significant improvement (19%-49% for sunflower meal and wheat, respectively) in comparison with those obtained using calibrations developed with linear methods.

Animal Feed↗
PubMed2006

Singularities affect dynamics of learning in neuromanifolds.

The parameter spaces of hierarchical systems such as multilayer perceptrons include singularities due to the symmetry and degeneration of hidden units. A parameter space forms a geometrical manifold, called the neuromanifold in the case of neural networks. Such a model is identified with a statistical model, and a Riemannian metric is given by the Fisher information matrix. However, the matrix degenerates at singularities. Such a singular structure is ubiquitous not only in multilayer perceptrons but also in the gaussian mixture probability densities, ARMA time-series model, and many other cases. The standard statistical paradigm of the Cramér-Rao theorem does not hold, and the singularity gives rise to strange behaviors in parameter estimation, hypothesis testing, Bayesian inference, model selection, and in particular, the dynamics of learning from examples. Prevailing theories so far have not paid much attention to the problem caused by singularity, relying only on ordinary statistical theories developed for regular (nonsingular) models. Only recently have researchers remarked on the effects of singularity, and theories are now being developed. This article gives an overview of the phenomena caused by the singularities of statistical manifolds related to multilayer perceptrons and gaussian mixtures. We demonstrate our recent results on these problems. Simple toy models are also used to show explicit solutions. We explain that the maximum likelihood estimator is no longer subject to the gaussian distribution even asymptotically, because the Fisher information matrix degenerates, that the model selection criteria such as AIC, BIC, and MDL fail to hold in these models, that a smooth Bayesian prior becomes singular in such models, and that the trajectories of dynamics of learning are strongly affected by the singularity, causing plateaus or slow manifolds in the parameter space. The natural gradient method is shown to perform well because it takes the singular geometrical structure into account. The generalization error and the training error are studied in some examples.

Journal Article↗
PubMed1998

Hebbian learning in parallel and modular memories.

Many cognitive and sensorimotor functions in the brain involve parallel and modular memory subsystems that are adapted by activity-dependent Hebbian synaptic plasticity. This is in contrast to the multilayer perceptron model of supervised learning where sensory information is presumed to be integrated by a common pool of hidden units through backpropagation learning. Here we show that Hebbian learning in parallel and modular memories is more advantageous than backpropagation learning in lumped memories in two respects: it is computationally much more efficient and structurally much simpler to implement with biological neurons. Accordingly, we propose a more biologically relevant neural network model, called a tree-like perceptron, which is a simple modification of the multilayer perceptron model to account for the general neural architecture, neuronal specificity, and synaptic learning rule in the brain. The model features a parallel and modular architecture in which adaptation of the input-to-hidden connection follows either a Hebbian or anti-Hebbian rule depending on whether the hidden units are excitatory or inhibitory, respectively. The proposed parallel and modular architecture and implicit interplay between the types of synaptic plasticity and neuronal specificity are exhibited by some neocortical and cerebellar systems.

Cerebellum↗