Implementing wavelet transform/mixture of experts network for analysis of electrocardiogram beats

Abstract: Mixture of experts (ME) is a modular neural network architecture for supervised learning. This paper illustrates the use of the ME network structure to guide model selection for classification of electrocardiogram (ECG) beats. The expectation maximization algorithm is used for training the ME so that the learning process is decoupled in a manner that fits well with the modular structure. The ECG signals were decomposed into time–frequency representations using discrete wavelet transforms and statistical features were calculated to depict their distribution. The ME network structure was implemented for ECG beats classification using the statistical features as inputs. To improve classification accuracy, the outputs of expert networks were combined by a gating network simultaneously trained in order to stochastically select the expert that is performing the best at solving the problem. Five types of ECG beats (normal beat, congestive heart failure beat, ventricular tachyarrhythmia beat, atrial fibrillation beat, partial epilepsy beat) obtained from the Physiobank database were classified with an accuracy of 96.89% by the ME network structure. The ME network structure achieved accuracy rates which were higher than those of the stand-alone neural network models.     

Recurrent neural networks employing Lyapunov exponents for analysis of doppler ultrasound signals

The implementation of recurrent neural networks (RNNs) with the Lyapunov exponents for Doppler ultrasound signals classification is presented. This study is based on the consideration that Doppler ultrasound signals are chaotic signals. This consideration was tested successfully using the nonlinear dynamics tools, like the computation of Lyapunov exponents. Decision making was performed in two stages: computation of Lyapunov exponents as representative features of the Doppler ultrasound signals and classification using the RNNs trained on the extracted features. The present research demonstrated that the Lyapunov exponents are the features which well represent the Doppler ultrasound signals and the RNNs trained on these features achieved high classification accuracies.     

Alendronic acid bearing ketone-formaldehyde resin and clay nanocomposites for fire-retardant polyurethanes

Cyclohexanone-formaldehyde resin and acetophenone-formaldehyde resin were in situ modified with sodium alendronate and the nanocomposites of cyclohexanone-formaldehyde resin with sepiolite and montmorillonite by adding the nanoclay into the resin formation media. The analysis of nanocomposites of nanoclay in the resin were carried out with X-ray diffraction. Exfoliated nanocomposites containing as much as 5 wt% sepiolite were obtained while the nanocomposites of montmorillonite were intercalated. The nanocomposites of alendronic acid-modified ketone-formaldehyde resin containing as much as 20 wt% sepiolite or 10 wt% montmorillonite were successfully produced in situ. Fourier transform infrared and nuclear magnetic resonance spectroscopies (NMR) were used for the structural characterization of the modified resins. The alendronic acid modified resins and their nanocomposites with the clays were used to produce reactive non-toxic fire retardant rigid polyurethane foam. The polyurethane foam containing alendronic acid-modified acetophenone-formaldehyde resin had a residue of 20% in the thermogravimetric analysis and was HB-grade in the horizontal fire test (HB) and self-extinguished in about 3–5 s during the test. Polyurethane foams containing alendronic acid and the nanoclay with the limiting oxygen index value of 20.8 were achieved.     

Real-Time Detection of Sanitary Sewer Overflows Using Neural Networks and Time Series Analysis

Sanitary sewer overflows (SSOs) are becoming of increasing concern as a health risk. Utilities and regulators have taken preventive measures but many overflows still occur and are not identifiable, especially in access-challenged locations. Several mathematical approaches are presented for detecting if a disruption in the system is impending or occurring based on measurements at one or more locations in the system. Time series analysis and neural networks are used as prediction tools for expected depths and flows for single measurement locations and a neural network is developed for a multiple monitor system. Control limit theory is applied in all cases for identifying significant deviations of measured values from the expected values that suggest a SSO is occurring. Data from Pima County Wastewater Management’s monitoring system are used in two case studies.     

Evaluation of call blocking probabilities in LEO satellite networks

In this paper, we present a new method for calculating call blocking probabilities (CBPs) in a low Earth orbit satellite network that carries voice calls. The calculation of the CBPs uses the Erlang‐B formula, but the traffic intensity has been modified to take into the time and location in which the calls are made. Copyright © 2009 John Wiley & Sons, Ltd.     

An ethanol electrooxidation study on carbon-supported Pt-Ru nanoparticles for direct ethanol fuel cells

Carbon supported Pt-Ru catalysts were prepared at varying Pt:Ru ratios by polyol method. The crystallite sizes of these catalysts were determined by X-ray diffraction technique. The specific surface areas of these catalysts were also defined by Brunauer–Emmett–Teller technique. Cyclic voltammetry, chronoamperometry, and electrochemical impedance spectroscopy measurements were conducted on these carbon supported Pt-Ru catalysts to investigate the effect of ruthenium on the ethanol electrooxidation kinetics. Results indicated that Pt-Ru (25:1) catalyst showed the best ethanol electrooxidation activity. In conclusion, ethanol electrooxidation mechanism was proposed.     

Changes in neighborhood poverty from 1990 to 2000 and youth's problem behaviors.

This study used data from the Project on Human Development in Chicago Neighborhoods, a multilevel, longitudinal study of children sampled from 80 diverse neighborhoods, to explore associations among changes in neighborhood poverty from 1990 to 2000 and changes in youth's internalizing problems and property and violent offenses over 6 years (N = 3,324; mean age across waves = 12.6 years). After accounting for a host of background characteristics and weighting for the propensity to stay in the original sampled neighborhood, results indicated that neighborhood poverty dynamics were unfavorably linked to boys' problem behaviors. In high-poverty (>30% in 1990) neighborhoods, decreasing poverty was associated with boys' greater internalizing problems and higher probability of increasing in violent behavior than stable neighborhood poverty. In moderate-poverty (20%–30% in 1990) neighborhoods, boys in neighborhoods that got poorer had more internalizing problems than boys in stably poor neighborhoods. Likewise, in low-poverty (     

Poly(3-hydroxyalkanoate)s: Diversification and biomedical applications: A state of the art review

Biomaterials have played an important role in the treatment of disease and the improvement of health care. Synthetic and naturally occurring biodegradable and biocompatible polymers have been used as biomaterials. Polyhydroxyalkanoates (PHAs) are promising materials for biomedical applications because they are biodegradable, non-toxic and biocompatible. We will shortly summarize the modification reactions, which include functionalization and grafting reactions, to improve the mechanical, thermal and hydrophilic properties of PHAs. The use of the modified PHAs in numerous biomedical applications, such as sutures, cardiovascular patches, wound dressings, scaffolds in tissue engineering, tissue repair/regeneration devices, drug carriers will be discussed in this review.     

A two-stage bid-price control for make-to-order revenue management

Capacity control implementations in make-to-order (MTO) revenue management typically are based on bid-prices, which are used to approximate the opportunity cost of accepting a customer request. However, in the face of stochastic demand, this approximation becomes less accurate and the performance of bid-prices may deteriorate. To address this problem, we examine the informational dynamics inherent in MTO capacity control problems and propose a two-stage capacity control approach based on bid-price updates. Updating is realized with neural networks, which are applied to adjust the selection criteria during the booking period with respect to online demand information. Not only is the resulting contribution margin positively influenced by the update, but also the downside risk of performing worse than a naive first-come-first-served policy. Results from computational experiments show that the proposed approach dominates traditional revenue management methods like randomized linear programming with and without resolving in expected contribution margin as well as in risk.