Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “Neural network”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 1,567 records · Page 87Linked to original sources
PubMedDate not supplied

On neural network techniques in the secure management of communication systems through improving and quality assessing pseudorandom stream generators.

Random components play an especially important role in the management of secure communication systems, with emphasis on the key management of cryptographic protocols. For this reason, the existence of strong pseudo random number generators is highly required. This paper presents novel techniques, which rely on Artificial Neural Network (ANN) architectures, to strengthen traditional generators such as IDEA and ANSI X.9 based on 3DES and IDEA. Additionally, this paper proposes a non-linear test method for the quality assessment of the required non-predictability property, which relies on feedforward neural networks. This non-predictability test method along with commonly used empirical tests based on statistics is proposed as a methodology for quality assessing strong pseudorandom stream generators. By means of this methodology, traditional and Neural Network based pseudorandom stream generators are evaluated. The results show that the proposed generators behave significantly better than the traditional ones, in particular, in terms of non-predictability.

Computer Communication Networks↗
PubMed1995

Dynamical recurrent neural networks--towards environmental time series prediction.

Dynamical Recurrent Neural Networks (DRNN) (Aussem 1995a) are a class of fully recurrent networks obtained by modeling synapses as autoregressive filters. By virtue of their internal dynamic, these networks approximate the underlying law governing the time series by a system of nonlinear difference equations of internal variables. They therefore provide history-sensitive forecasts without having to be explicitly fed with external memory. The model is trained by a local and recursive error propagation algorithm called temporal-recurrent-backpropagation. The efficiency of the procedure benefits from the exponential decay of the gradient terms backpropagated through the adjoint network. We assess the predictive ability of the DRNN model with meterological and astronomical time series recorded around the candidate observation sites for the future VLT telescope. The hope is that reliable environmental forecasts provided with the model will allow the modern telescopes to be preset, a few hours in advance, in the most suited instrumental mode. In this perspective, the model is first appraised on precipitation measurements with traditional nonlinear AR and ARMA techniques using feedforward networks. Then we tackle a complex problem, namely the prediction of astronomical seeing, known to be a very erratic time series. A fuzzy coding approach is used to reduce the complexity of the underlying laws governing the seeing. Then, a fuzzy correspondence analysis is carried out to explore the internal relationships in the data. Based on a carefully selected set of meteorological variables at the same time-point, a nonlinear multiple regression, termed nowcasting (Murtagh et al. 1993, 1995), is carried out on the fuzzily coded seeing records. The DRNN is shown to outperform the fuzzy k-nearest neighbors method.

Environmental Monitoring↗
PubMed1998

Analysis of quantitative EEG with artificial neural networks and discriminant analysis--a methodological comparison.

Artificial neural networks (ANN) are widely used to solve problems of differentiating between groups. However, serious comparisons of this method with the traditional procedure for such tasks (discriminant analysis) are rare. Discussing the results of both methods with the example of highly topical data, we try to demonstrate advantages and drawbacks of both methods. For this purpose, quantitative EEGs of 78 alcoholics were investigated in order to determine whether it is possible to predict relapse of these patients at the beginning of treatment. ANN software is available in Kassel (Institute for Computer Sciences and Mathematics).

Alcoholism↗
PubMed2003

Optimization of the separation of organic explosives by capillary electrophoresis with artificial neural networks.

The separation of 12 explosives by capillary electrophoresis was optimized with the aid of artificial neural networks (ANNs). The selectivity of the separation was manipulated by varying the concentration of sodium dodecyl sulfate (SDS) and the pH of the electrolyte, while maintaining the buffer concentration at 10 mM borate. The concentration of SDS and the electrolyte pH were used as input variables and the mobility of the explosives were used as output variables for the ANN. In total, eight experiments were performed based on a factorial design to train a variety of artificial neural network architectures. A further three experiments were required to train ANN architectures to adequately model the experimental space. A product resolution response surface was constructed based on the predicted mobilities of the best performing ANN. This response surface pointed to two optima; pH 9.0-9.1 and 60-65 mM SDS, and pH 8.4-8.6 and 50-60 mM SDS. Separation of all 12 explosives was achieved at the second optimum. The separation was further improved by changing the capillary to an extended cell detection window and reducing the diameter of the capillary from 75 microm to 50 microm. This provided a more efficient separation without compromising detection sensitivity.

Aniline Compounds↗
PubMed2001

[Efficiency of artificial neural networks for prediction of regional lymph node metastasis].

A retrospective study of 90 patients suffering from laryngeal carcinoma who were surgically treated in the Department of Otolaryngology, Collegium Medicum, Jagiellonian University, Cracow, Poland, has been performed. In 26 patients metastases in regional lymph nodes has been found. Various clinical, histopathological and kariometric factors have been subject to analysis. The artificial neural networks techniques have been used for prediction of presence of metastases in lymph nodes. The artificial neural networks exhibit an ability to generalize, i.e. they can learn from exemplary sets of data. Therefore the patients data have been divided at random into two sets forming a learning set (70 patients) and a test set (20 patients). A series of simulations have been carried out, using networks of various architectures and experimentally selected learning parameters, giving the prediction results between 71% and 96% of correct recognitions. The best prediction results have been achieved for the network of triple layer perceptron type including 6 neurons in the hidden layer. The evaluation of factors related to the presence of metastases in lymph nodes has been also carried out using logistic regression method for several groups of patients including various numbers of cases. The data sets of the remaining patients have been used for testing the reliability of that statistical method, giving between 56% and 64% of correct recognitions. Therefore the possibilities of prediction of metastases presence in lymph nodes are better for the artificial neural networks method (up to 96% of correct classifications) i.m. comparison to the logistic regression method (only up to 64% of correct classifications for the same data sets).

Humans↗
PubMed2004

Artificial neural networks for predictive modeling in prostate cancer.

Artificial neural networks (ANNs) represent a relatively new methodology for predictive modeling in medicine. ANNs, a form of artificial intelligence loosely based on the brain, have a demonstrated ability to learn complex and subtle relationships between variables in medical applications. In contrast with traditional statistical techniques, ANNs are capable of automatically resolving these relationships without the need for a priori assumptions about the nature of the interactions between variables. As with any technique, ANNs have limitations and potential drawbacks. This article provides an overview of the theoretical basis of ANNs, how they function, their strengths and limitations, and examples of how ANNs have been used to develop predictive models for the management of prostate cancer.

Diagnosis, Differential↗
PubMed2006

[Assessment of giant panda habitat based on integration of expert system and neural network].

To conserve giant panda effectively, it is important to understand the spatial pattern and temporal change of its habitat. Mapping is an effective approach for wildlife habitat evaluation and monitoring. The application of recently developed artificial intelligence tools, including expert systems and neural networks, could integrate qualitative and quantitative information for modeling complex systems, and built the information into a GIS, which could be helpful for giant panda habitat mapping. This study built a mapping approach for giant panda habitat mapping, which integrated expert system and neural network classifiers (ESNNC), and used multi-type data within GIS. The giant panda habitat types and their suitability were mapped by ESNNC. The results showed that the habitat types and their suitability in Foping Nature Reserve were assessed with a higher accuracy (> 80 %) by ESNNC, compared with non-integrated classifiers, i. e., expert system, neural network, and maximum likelihood. Z-statistic test showed that ESNNC was significantly better than the other three non-integrated classifiers. It was recommended that the integrated approach could be widely applied into wildlife habitat assessment.

Animals↗
PubMed2004

Selecting compounds for focused screening using linear discriminant analysis and artificial neural networks.

Linear discriminant analysis and a committee of neural networks have been applied to recognise compounds that act at biological targets belonging to a specific gene family, protein kinases. The MDDR database was used to provide compounds targeted against this family and sets of randomly selected molecules. BCUT parameters were employed as input descriptors that encode structural properties and information relevant to ligand-receptor interactions. The technique was applied to purchasing compounds from external suppliers. These compounds achieved hit rates on a par with those achieved using known actives for related targets when tested for the ability to inhibit kinases at a single concentration. This approach is intended as one of a series of filters in the selection of screening candidates, compound purchases and the application of synthetic priorities to combinatorial libraries.

Databases, Factual↗
PubMed2005

Artificial neural network computer prediction of ovarian malignancy in women with adnexal masses.

OBJECTIVE: Assessment of the usefulness of a neural model to predict which ovarian tumors are malignant. METHOD: Age, menopausal status, body mass index, grayscale and Doppler ultrasonographic features, as well as levels of specific markers (CA 125, tissue polypeptide specific antigen) were examined in 686 women with adnexal masses. The probability of malignancy was calculated using an artificial neural network software and the diagnostic efficiency of the received model was estimated using a receiver-operating characteristics (ROC) curve. RESULT: Of the 686 women, 431 (62.8%) had a benign and 255 (37.2%) had a malignant ovarian tumor. The significant malignancy predictors are age, menopausal status, maximum tumor diameter, internal wall structure of tumor, presence of septa and/or solid elements, tumor location, location of vessels, and blood flow indexes. The best network provided 96.0% sensitivity and 97.7% specificity. The area under the curve for the received model was 0.9716. CONCLUSIONS: An artificial neural network model based on clinical and ultrasonographic data allows to calculate the probability of tumor malignancy.

Adnexal Diseases↗
PubMed1998

Evolution of artificial neural network architecture: prediction of depression after mania.

Artificial neural networks (ANNs) are compared to standard statistical methods for outcome prediction in biomedical problems. A general method for using genetic algorithms to "evolve" ANN architecture (EANN) is presented. Accuracy of logistic regression, a fully interconnected ANN, and an EANN for predicting depression after mania are examined. All methods showed very good agreement (training set accuracy, chi-square all p < 0.01). However, significant differences were found for stability (test set accuracy); logistic regression being the most unstable and EANN being significantly more stable than a fully interconnected ANN (McNemar p < 0.01). We conclude that the EANN method enhances ANN stability. This approach may have particular relevance for biomedical prediction problems, such as predicting depression after mania.

Algorithms↗
PubMed2006

A hybrid forward algorithm for RBF neural network construction.

This paper proposes a novel hybrid forward algorithm (HFA) for the construction of radial basis function (RBF) neural networks with tunable nodes. The main objective is to efficiently and effectively produce a parsimonious RBF neural network that generalizes well. In this study, it is achieved through simultaneous network structure determination and parameter optimization on the continuous parameter space. This is a mixed integer hard problem and the proposed HFA tackles this problem using an integrated analytic framework, leading to significantly improved network performance and reduced memory usage for the network construction. The computational complexity analysis confirms the efficiency of the proposed algorithm, and the simulation results demonstrate its effectiveness.

Algorithms↗
PubMedDate not supplied

Prediction of occurrences of diverse chemical classes in the Asteraceae through artificial neural networks.

The training and the application of a neural network system for the prediction of occurrences of secondary metabolites belonging to diverse chemical classes in the Asteraceae is described. From a database containing about 604 genera and 28,000 occurrences of secondary metabolites in the plant family, information was collected encompassing nine chemical classes and their respective occurrences for training of a multi-layer net using the back-propagation algorithm. The net supplied as output the presence or absence of the chemical classes as well as the number of compounds isolated from each taxon. The results provided by the net from the presence or absence of a chemical class showed a 89% hit rate; by excluding triterpenes from the analysis, only 5% of the genera studied exhibited errors greater than 10%.

Asteraceae↗
PubMed2003

Calculation of PID controller parameters by using a fuzzy neural network.

In this paper, we use the fuzzy neural network (FNN) to develop a formula for designing the proportional-integral-derivative (PID) controller. This PID controller satisfies the criteria of minimum integrated absolute error (IAE) and maximum of sensitivity (Ms). The FNN system is used to identify the relationship between plant model and controller parameters based on IAE and Ms. To derive the tuning rule, the dominant pole assignment method is applied to simplify our optimization processes. Therefore, the FNN system is used to automatically tune the PID controller for different system parameters so that neither theoretical methods nor numerical methods need be used. Moreover, the FNN-based formula can modify the controller to meet our specification when the system model changes. A simulation result for applying to the motor position control problem is given to demonstrate the effectiveness of our approach.

Algorithms↗
PubMed2002

Space-filling curves and Kolmogorov superposition-based neural networks.

Kolmogorov superpositions and Hecht-Nielsen's neural network based on them are dimension reducing. This dimension reduction can be understood in terms of space-filling curves that characterize Kolmogorov's functions, and the subject of this paper is the construction of such a curve. We construct a space-filling curve with Lebesgue measure 1 in the unit square, [0,1]2, with approximating curves lambda(k), k = 1,2,3,..., each with 10(2k) rational nodal-points whose order is determined for each k by the linear order of their image-points under a nomographic function y = alpha1psi(x1) + alpha2psi(x2) that is the basis of a computable version of the Kolmogorov superpositions in two dimensions. The function psi:[0,1] --> [0,1] is continuous and monotonic increasing, and alpha1, alpha2 are suitable constants. The curves lambda(k) are composed of families of disjoint closed squares of diminishing diameters and connecting joins of diminishing lengths as k --> infinity.

Neural Networks, Computer↗
PubMed2002

Comparison of Artificial Neural Network (ANN) Model Development Methods for Prediction of Macroinvertebrate Communities in the Zwalm River Basin in Flanders, Belgium.

Modelling has become an interesting tool to support decision making in water management. River ecosystem modelling methods have improved substantially during recent years. New concepts, such as artificial neural networks, fuzzy logic, evolutionary algorithms, chaos and fractals, cellular automata, etc., are being more commonly used to analyse ecosystem databases and to make predictions for river management purposes. In this context, artificial neural networks were applied to predict macroinvertebrate communities in the Zwalm River basin (Flanders, Belgium). Structural characteristics (meandering, substrate type, flow velocity) and physical and chemical variables (dissolved oxygen, pH) were used as predictive variables to predict the presence or absence of macroinvertebrate taxa in the headwaters and brooks of the Zwalm River basin. Special interest was paid to the frequency of occurrence of the taxa as well as the selection of the predictors and variables to be predicted on the prediction reliability of the developed models. Sensitivity analyses allowed us to study the impact of the predictive variables on the prediction of presence or absence of macroinvertebrate taxa and to define which variables are the most influential in determining the neural network outputs.

Amphipoda↗
PubMed1994

Rapid identification of species within the Mycobacterium tuberculosis complex by artificial neural network analysis of pyrolysis mass spectra.

An artificial neural network (ANN) was trained to distinguish between Mycobacterium tuberculosis and M. bovis with averaged pyrolysis mass spectra from duplicate subcultures of four strains of each of these species, each pyrolysed in triplicate. Once trained, the ANN was interrogated with spectrum data from the original organisms (the "training set") and from 26 other mycobacterial isolates (the "challenge set") of the M. tuberculosis complex (MTBC). Eight strains of M. bovis and 13 of M. tuberculosis, whether sensitive or variously resistant to antituberculosis drugs, were identified in agreement with conventional identification. Four strains of "M. africanum" were identified as M. bovis. Of two atypical M. tuberculosis strains from South India, one was identified as M. tuberculosis and the other as M. bovis. Six strains of BCG proved heterogeneous; two gave equivocal identifications, three were identified as M. bovis and one was identified as M. tuberculosis.

Humans↗
PubMed1998

Using neural networks for prediction of the subcellular location of proteins.

Neural networks have been trained to predict the subcellular location of proteins in prokaryotic or eukaryotic cells from their amino acid composition. For three possible subcellular locations in prokaryotic organisms a prediction accuracy of 81% can be achieved. Assigning a reliability index, 33% of the predictions can be made with an accuracy of 91%. For eukaryotic proteins (excluding plant sequences) an overall prediction accuracy of 66% for four locations was achieved, with 33% of the sequences being predicted with an accuracy of 82% or better. With the subcellular location restricting a protein's possible function, this method should be a useful tool for the systematic analysis of genome data and is available via a server on the world wide web.

Amino Acids↗
PubMed1997

A comparison of quantitative structure-activity relationships for the effect of benzoic and cinnamic acids on Listeria monocytogenes using multiple linear regression, artificial neural network and fuzzy systems.

The ability of artificial neural networks (ANN), fuzzy systems (FS) and multiple linear regression (MLR) to fit the biological activity surface describing the inhibition of Listeria monocytogenes by benzoic and cinnamic acid derivatives was compared. MLR and ANN were also compared for their ability to select the properties that best describe the biological activity of the compounds. The criteria used for comparing surface fits of all models were the coefficient of determination r2 and the standard deviation of the error, s(e). The ANN method gave a better correlation, r2 = 0.96, compared with either MLR, r2 = 0.81, or FS, r2 = 0.92, and also a lower standard error, possibly indicating non-linearity in the data. The ANN was shown to generalize better than MLR using the leave-one-out method. The ANN selection algorithm for the selection of the parameters that contributed most to the biological activity of the phenols (log K and pKa) agreed with the selected parameters of the MLR system.

Benzaldehydes↗