Simulated annealing technique for fast learning of SOM networks

The Self-Organizing Map (SOM) is a popular unsupervised neural network able to provide effective clustering and data visualization for multidimensional input datasets. In this paper, we present an application of the simulated annealing procedure to the SOM learning algorithm with the aim to obtain a fast learning and better performances in terms of quantization error. The proposed learning algorithm is called Fast Learning Self-Organized Map, and it does not affect the easiness of the basic learning algorithm of the standard SOM. The proposed learning algorithm also improves the quality of resulting maps by providing better clustering quality and topology preservation of input multi-dimensional data. Several experiments are used to compare the proposed approach with the original algorithm and some of its modification and speed-up techniques.

A combined engineering and statistical model of UK domestic appliance electrical load profiles

The development of a combined engineering and statistical Artificial Neural Network model of UK domestic appliance load profiles is presented. The model uses diary-style appliance use data and a survey questionnaire collected from 51 suburban households and 46 rural households during the summer of 2010 and2011 respectively. It also incorporates measured energy data and is sensitive to socioeconomic, physical dwelling and temperature variables. A prototype model is constructed in MATLAB using a two layer feed forward network with back propagation training which has a 12:10:24 architecture. Model outputs include appliance load profiles which can be applied to the fields of energy planning (microrenewables and smart grids), building simulation tools and energy policy.

Predicting metal-binding site residues in low-resolution structural models

The accurate prediction of the biochemical function of a protein is becoming increasingly important, given the unprecedented growth of both structural and sequence databanks. Consequently, computational methods are required to analyse such data in an automated manner to ensure genomes are annotated accurately. Protein structure prediction methods, for example, are capable of generating approximate structural models on a genome-wide scale. However, the detection of functionally important regions in such crude models, as well as structural genomics targets, remains an extremely important problem. The method described in the current study, MetSite, represents a fully automatic approach for the detection of metal-binding residue clusters applicable to protein models of moderate quality. The method involves using sequence profile information in combination with approximate structural data. Several neural network classifiers are shown to be able to distinguish metal sites from non-sites with a mean accuracy of 94.5%. The method was demonstrated to identify metal-binding sites correctly in LiveBench targets where no obvious metal-binding sequence motifs were detectable using InterPro. Accurate detection of metal sites was shown to be feasible for low-resolution predicted structures generated using mGenTHREADER where no side-chain information was available. High-scoring predictions were observed for a recently solved hypothetical protein from Haemophilus influenzae, indicating a putative metal-binding site.

Automatic prediction of functional site regions in low-resolution protein structures

World-wide structural genomics initiatives are rapidly accumulating structures for which limited functional information is available. Additionally, state-of-the art structural prediction programs are now capable of generating at least low resolution structural models of target proteins. Accurate detection and classification of functional sites within both solved and modelled protein structures therefore represents an important challenge. We present a fully automatic site detection method, FuncSite, that uses neural network classifiers to predict the location and type of functionally important sites in protein structures. The method is designed primarily to require only backbone residue positions without the need for specific side-chain atoms to be present. In order to highlight effective site detection in low resolution structural models FuncSite was used to screen model proteins generated using mGenTHREADER on a set of newly released structures. We found effective metal site detection even for moderate quality protein models illustrating the robustness of the method.

Improvement of the GenTHREADER method for genomic fold recognition

Motivation: In order to enhance genome annotation, the fully automatic fold recognition method GenTHREADER has been improved and benchmarked. The previous version of GenTHREADER consisted of a simple neural network which was trained to combine sequence alignment score, length information and energy potentials derived from threading into a single score representing the relationship between two proteins, as designated by CATH. The improved version incorporates PSI-BLAST searches, which have been jumpstarted with structural alignment profiles from FSSP, and now also makes use of PSIPRED predicted secondary structure and bi-directional scoring in order to calculate the final alignment score. Pairwise potentials and solvation potentials are calculated from the given sequence alignment which are then used as inputs to a multi-layer, feed-forward neural network, along with the alignment score, alignment length and sequence length. The neural network has also been expanded to accommodate the secondary structure element alignment (SSEA) score as an extra input and it is now trained to learn the FSSP Z-score as a measurement of similarity between two proteins. Results: The improvements made to GenTHREADER increase the number of remote homologues that can be detected with a low error rate, implying higher reliability of score, whilst also increasing the quality of the models produced. We find that up to five times as many true positives can be detected with low error rate per query. Total MaxSub score is doubled at low false positive rates using the improved method.

System identification of Wiener systems with B-spline functions using De Boor recursion

In this article a simple and effective algorithm is introduced for the system identification of the Wiener system using observational input/output data. The nonlinear static function in the Wiener system is modelled using a B-spline neural network. The Gauss–Newton algorithm is combined with De Boor algorithm (both curve and the first order derivatives) for the parameter estimation of the Wiener model, together with the use of a parameter initialisation scheme. Numerical examples are utilised to demonstrate the efficacy of the proposed approach.

Modelling and inverting complex-valued Wiener systems

We develop a complex-valued (CV) B-spline neural network approach for efficient identification and inversion of CV Wiener systems. The CV nonlinear static function in the Wiener system is represented using the tensor product of two univariate B-spline neural networks. With the aid of a least squares parameter initialisation, the Gauss-Newton algorithm effectively estimates the model parameters that include the CV linear dynamic model coefficients and B-spline neural network weights. The identification algorithm naturally incorporates the efficient De Boor algorithm with both the B-spline curve and first order derivative recursions. An accurate inverse of the CV Wiener system is then obtained, in which the inverse of the CV nonlinear static function of the Wiener system is calculated efficiently using the Gaussian-Newton algorithm based on the estimated B-spline neural network model, with the aid of the De Boor recursions. The effectiveness of our approach for identification and inversion of CV Wiener systems is demonstrated using the application of digital predistorter design for high power amplifiers with memory

FunFOLDQA: a quality assessment tool for protein-ligand binding site residue predictions

The estimation of prediction quality is important because without quality measures, it is difficult to determine the usefulness of a prediction. Currently, methods for ligand binding site residue predictions are assessed in the function prediction category of the biennial Critical Assessment of Techniques for Protein Structure Prediction (CASP) experiment, utilizing the Matthews Correlation Coefficient (MCC) and Binding-site Distance Test (BDT) metrics. However, the assessment of ligand binding site predictions using such metrics requires the availability of solved structures with bound ligands. Thus, we have developed a ligand binding site quality assessment tool, FunFOLDQA, which utilizes protein feature analysis to predict ligand binding site quality prior to the experimental solution of the protein structures and their ligand interactions. The FunFOLDQA feature scores were combined using: simple linear combinations, multiple linear regression and a neural network. The neural network produced significantly better results for correlations to both the MCC and BDT scores, according to Kendall’s τ, Spearman’s ρ and Pearson’s r correlation coefficients, when tested on both the CASP8 and CASP9 datasets. The neural network also produced the largest Area Under the Curve score (AUC) when Receiver Operator Characteristic (ROC) analysis was undertaken for the CASP8 dataset. Furthermore, the FunFOLDQA algorithm incorporating the neural network, is shown to add value to FunFOLD, when both methods are employed in combination. This results in a statistically significant improvement over all of the best server methods, the FunFOLD method (6.43%), and one of the top manual groups (FN293) tested on the CASP8 dataset. The FunFOLDQA method was also found to be competitive with the top server methods when tested on the CASP9 dataset. To the best of our knowledge, FunFOLDQA is the first attempt to develop a method that can be used to assess ligand binding site prediction quality, in the absence of experimental data.

What determines the out-of-home placement of children in the USA?

Using NCANDS data of US child maltreatment reports for 2009, logistic regression, probit analysis, discriminant analysis and an artificial neural network are used to determine the factors which explain the decision to place a child in out-of-home care. As well as developing a new model for 2009, a previous study using 2005 data is replicated. While there are many small differences, the four estimation techniques give broadly the same results, demonstrating the robustness of the results. Similarly, apart from age and sexual abuse, the 2005 and 2009 results are roughly similar. For 2009, child characteristics (particularly child emotional problems) are more important than the nature of the abuse and the situation of the household; while caregiver characteristics are the least important. All these models have low explanatory power.

Parkinsonian tremor identification with multiple local field potential feature classification

This paper explores the development of multi-feature classification techniques used to identify tremor-related characteristics in the Parkinsonian patient. Local field potentials were recorded from the subthalamic nucleus and the globus pallidus internus of eight Parkinsonian patients through the implanted electrodes of a Deep brain stimulation (DBS) device prior to device internalization. A range of signal processing techniques were evaluated with respect to their tremor detection capability and used as inputs in a multi-feature neural network classifier to identify the activity of Parkinsonian tremor. The results of this study show that a trained multi-feature neural network is able, under certain conditions, to achieve excellent detection accuracy on patients unseen during training. Overall the tremor detection accuracy was mixed, although an accuracy of over 86% was achieved in four out of the eight patients.

Combined neurostimulation and neuroimaging in cognitive neuroscience: past, present, and future

Modern neurostimulation approaches in humans provide controlled inputs into the operations of cortical regions, with highly specific behavioral consequences. This enables causal structure–function inferences, and in combination with neuroimaging, has provided novel insights into the basic mechanisms of action of neurostimulation on dis- tributed networks. For example,more recent work has established the capacity of transcranialmagnetic stimulation (TMS) to probe causal interregional influences, and their interaction with cognitive state changes. Combinations of neurostimulation and neuroimaging now face the challenge of integrating the known physiological effects of neu- rostimulationwith theoretical and biologicalmodels of cognition, for example,when theoretical stalemates between opposing cognitive theories need to be resolved. This will be driven by novel developments, including biologically informedcomputational network analyses for predicting the impactofneurostimulationonbrainnetworks, as well as novel neuroimaging and neurostimulation techniques. Such future developments may offer an expanded set of tools withwhich to investigate structure–function relationships, and to formulate and reconceptualize testable hypotheses about complex neural network interactions and their causal roles in cognition

A modified PNN algorithm with optimal PD modeling using the orthogonal least squares method

In this paper a modified algorithm is suggested for developing polynomial neural network (PNN) models. Optimal partial description (PD) modeling is introduced at each layer of the PNN expansion, a task accomplished using the orthogonal least squares (OLS) method. Based on the initial PD models determined by the polynomial order and the number of PD inputs, OLS selects the most significant regressor terms reducing the output error variance. The method produces PNN models exhibiting a high level of accuracy and superior generalization capabilities. Additionally, parsimonious models are obtained comprising a considerably smaller number of parameters compared to the ones generated by means of the conventional PNN algorithm. Three benchmark examples are elaborated, including modeling of the gas furnace process as well as the iris and wine classification problems. Extensive simulation results and comparison with other methods in the literature, demonstrate the effectiveness of the suggested modeling approach.

Predicting stock index volatility: can market volume help?

This paper explores a number of statistical models for predicting the daily stock return volatility of an aggregate of all stocks traded on the NYSE. An application of linear and non-linear Granger causality tests highlights evidence of bidirectional causality, although the relationship is stronger from volatility to volume than the other way around. The out-of-sample forecasting performance of various linear, GARCH, EGARCH, GJR and neural network models of volatility are evaluated and compared. The models are also augmented by the addition of a measure of lagged volume to form more general ex-ante forecasting models. The results indicate that augmenting models of volatility with measures of lagged volume leads only to very modest improvements, if any, in forecasting performance.

On the application of optimal wavelet filter banks for ECG signal classification

This paper discusses ECG signal classification after parametrizing the ECG waveforms in the wavelet domain. Signal decomposition using perfect reconstruction quadrature mirror filter banks can provide a very parsimonious representation of ECG signals. In the current work, the filter parameters are adjusted by a numerical optimization algorithm in order to minimize a cost function associated to the filter cut-off sharpness. The goal consists of achieving a better compromise between frequency selectivity and time resolution at each decomposition level than standard orthogonal filter banks such as those of the Daubechies and Coiflet families. Our aim is to optimally decompose the signals in the wavelet domain so that they can be subsequently used as inputs for training to a neural network classifier.

Nonlinear equalization of Hammerstein OFDM systems

A practical orthogonal frequency-division multiplexing (OFDM) system can generally be modelled by the Hammerstein system that includes the nonlinear distortion effects of the high power amplifier (HPA) at transmitter. In this contribution, we advocate a novel nonlinear equalization scheme for OFDM Hammerstein systems. We model the nonlinear HPA, which represents the static nonlinearity of the OFDM Hammerstein channel, by a B-spline neural network, and we develop a highly effective alternating least squares algorithm for estimating the parameters of the OFDM Hammerstein channel, including channel impulse response coefficients and the parameters of the B-spline model. Moreover, we also use another B-spline neural network to model the inversion of the HPA’s nonlinearity, and the parameters of this inverting B-spline model can easily be estimated using the standard least squares algorithm based on the pseudo training data obtained as a byproduct of the Hammerstein channel identification. Equalization of the OFDM Hammerstein channel can then be accomplished by the usual one-tap linear equalization as well as the inverse B-spline neural network model obtained. The effectiveness of our nonlinear equalization scheme for OFDM Hammerstein channels is demonstrated by simulation results.