The influence of ARIMA-GARCH parameters in feed forward neural networks prediction

The objective of this article is to find out the influence of the parameters of the ARIMA-GARCH models in the prediction of artificial neural networks (ANN) of the feed forward type, trained with the Levenberg-Marquardt algorithm, through Monte Carlo simulations. The paper presents a study of the relationship between ANN performance and ARIMA-GARCH model parameters, i.e. the fact that depending on the stationarity and other parameters of the time series, the ANN structure should be selected differently. Neural networks have been widely used to predict time series and their capacity for dealing with non-linearities is a normally outstanding advantage. However, the values of the parameters of the models of generalized autoregressive conditional heteroscedasticity have an influence on ANN prediction performance. The combination of the values of the GARCH parameters with the ARIMA autoregressive terms also implies in ANN performance variation. Combining the parameters of the ARIMA-GARCH models and changing the ANN`s topologies, we used the Theil inequality coefficient to measure the prediction of the feed forward ANN.

Automatic identification of terpenoid skeletons by feed-forward neural networks

Feed-forward neural networks (FFNNs) were used to predict the skeletal type of molecules belonging to six classes of terpenoids. A database that contains the (13)C NMR spectra of about 5000 compounds was used to train the FFNNs. An efficient representation of the spectra was designed and the constitution of the best FFNN input vector format resorted from an heuristic approach. The latter was derived from general considerations on terpenoid structures. (c) 2006 Elsevier B.V. All rights reserved.

Multi-antenna non-line-of-sight identification techniques for target localization in mobile ad-hoc networks

Target localization has a wide range of military and civilian applications in wireless mobile networks. Examples include battle-field surveillance, emergency 911 (E911), traffc alert, habitat monitoring, resource allocation, routing, and disaster mitigation. Basic localization techniques include time-of-arrival (TOA), direction-of-arrival (DOA) and received-signal strength (RSS) estimation. Techniques that are proposed based on TOA and DOA are very sensitive to the availability of Line-of-sight (LOS) which is the direct path between the transmitter and the receiver. If LOS is not available, TOA and DOA estimation errors create a large localization error. In order to reduce NLOS localization error, NLOS identifcation, mitigation, and localization techniques have been proposed. This research investigates NLOS identifcation for multiple antennas radio systems. The techniques proposed in the literature mainly use one antenna element to enable NLOS identifcation. When a single antenna is utilized, limited features of the wireless channel can be exploited to identify NLOS situations. However, in DOA-based wireless localization systems, multiple antenna elements are available. In addition, multiple antenna technology has been adopted in many widely used wireless systems such as wireless LAN 802.11n and WiMAX 802.16e which are good candidates for localization based services. In this work, the potential of spatial channel information for high performance NLOS identifcation is investigated. Considering narrowband multiple antenna wireless systems, two xvNLOS identifcation techniques are proposed. Here, the implementation of spatial correlation of channel coeffcients across antenna elements as a metric for NLOS identifcation is proposed. In order to obtain the spatial correlation, a new multi-input multi-output (MIMO) channel model based on rough surface theory is proposed. This model can be used to compute the spatial correlation between the antenna pair separated by any distance. In addition, a new NLOS identifcation technique that exploits the statistics of phase difference across two antenna elements is proposed. This technique assumes the phases received across two antenna elements are uncorrelated. This assumption is validated based on the well-known circular and elliptic scattering models. Next, it is proved that the channel Rician K-factor is a function of the phase difference variance. Exploiting Rician K-factor, techniques to identify NLOS scenarios are proposed. Considering wideband multiple antenna wireless systems which use MIMO-orthogonal frequency division multiplexing (OFDM) signaling, space-time-frequency channel correlation is exploited to attain NLOS identifcation in time-varying, frequency-selective and spaceselective radio channels. Novel NLOS identi?cation measures based on space, time and frequency channel correlation are proposed and their performances are evaluated. These measures represent a better NLOS identifcation performance compared to those that only use space, time or frequency.

A synthesis method for broad-band antenna impedance matching networks /

Vita.

A study of certain factors affecting computation times in designing optimal NOR networks by the implicit enumeration method using the feed-forward formulation /

Bibliography: p. 27-29.

Direct support and general support maintenance manual : antenna and microwave equipment including feed assembly AS-2941 ( )/FSC, (NSN 5895-00-639-4793), dehumidifier, desiccant, electric HD-988/G, (NSN 4440-01-018-0080) for satellite communication terminals AN/FSC-78(V) and AN/FSC-79.

Includes index.

Curvature-driven smoothing: a learning algorithm for feed-forward networks

The performance of feed-forward neural networks in real applications can be often be improved significantly if use is made of a-priori information. For interpolation problems this prior knowledge frequently includes smoothness requirements on the network mapping, and can be imposed by the addition to the error function of suitable regularization terms. The new error function, however, now depends on the derivatives of the network mapping, and so the standard back-propagation algorithm cannot be applied. In this paper, we derive a computationally efficient learning algorithm, for a feed-forward network of arbitrary topology, which can be used to minimize the new error function. Networks having a single hidden layer, for which the learning algorithm simplifies, are treated as a special case.

Regularization and complexity control in feed-forward networks

In this paper we consider four alternative approaches to complexity control in feed-forward networks based respectively on architecture selection, regularization, early stopping, and training with noise. We show that there are close similarities between these approaches and we argue that, for most practical applications, the technique of regularization should be the method of choice.

Topographic mappings and feed-forward neural networks

This thesis is a study of the generation of topographic mappings - dimension reducing transformations of data that preserve some element of geometric structure - with feed-forward neural networks. As an alternative to established methods, a transformational variant of Sammon's method is proposed, where the projection is effected by a radial basis function neural network. This approach is related to the statistical field of multidimensional scaling, and from that the concept of a 'subjective metric' is defined, which permits the exploitation of additional prior knowledge concerning the data in the mapping process. This then enables the generation of more appropriate feature spaces for the purposes of enhanced visualisation or subsequent classification. A comparison with established methods for feature extraction is given for data taken from the 1992 Research Assessment Exercise for higher educational institutions in the United Kingdom. This is a difficult high-dimensional dataset, and illustrates well the benefit of the new topographic technique. A generalisation of the proposed model is considered for implementation of the classical multidimensional scaling (¸mds}) routine. This is related to Oja's principal subspace neural network, whose learning rule is shown to descend the error surface of the proposed ¸mds model. Some of the technical issues concerning the design and training of topographic neural networks are investigated. It is shown that neural network models can be less sensitive to entrapment in the sub-optimal global minima that badly affect the standard Sammon algorithm, and tend to exhibit good generalisation as a result of implicit weight decay in the training process. It is further argued that for ideal structure retention, the network transformation should be perfectly smooth for all inter-data directions in input space. Finally, there is a critique of optimisation techniques for topographic mappings, and a new training algorithm is proposed. A convergence proof is given, and the method is shown to produce lower-error mappings more rapidly than previous algorithms.

The application of fibre Bragg grating networks as strain sensors and as phased array antenna true time delay elements

The fabrication of in-fibre Bragg gratings, and the application of arrays of such gratings as strain sensors and as true time delay elements for the control of phased array antennas is reported. Chirped period Bragg gratings were produced using the fibre deformation fabrication technique, with chirps of between 2.9nm and 17.3nm achieved. Arrays of 5mm and 2mm long uniform period Bragg gratings were fabricated using the inscription method, for use as true time delay elements,dissimilar wavefronts and their spectral characteristics recorded. The uniform period grating arrays were used to create minimum time delays of 9.09ps, 19.02ps and 31ps; making them suitable for controlling phased array antennas operating at RF frequencies of up to 3GHz, with 10° phase resolution. Four 4mm long chirped gratings were produced using the dissimilar wavefronts fabrication method, having chirps of 7nm, 12nm, 20nm and 30nm, and were used to create time delays of between 0.3ps and 59ps. Hence they are suitable for controlling phased array antennas at RF frequencies of up to 48GHz. The application of in fibre Bragg gratings as strain sensors within smart structure materials was investigated, with their sensitivity to applied strain and compression measured for both embedded and surface mounted uniform period and fibre Fabry-Perot filter gratings. A fibre Bragg grating sensor demultiplexing scheme based on a liquid crystal filled Fabry-Perot etalon tuneable transmission filter was proposed, successfully constructed and fully characterised. Three characteristics of the LCFP etalon were found to pose operational limitations to its application in a Bragg grating sensor system; most significantly, the resonance peak wavelength was highly (-2,77nm/°C) temperature dependent. Several methods for minimising this temperature sensitivity were investigated, but enjoyed only limited success. It was therefore concluded that this type (E7 filled) of LCFP etalon is unsuitable for use as a Bragg grating sensor demultiplexing element.

Effect of the base station antenna beam tilting on energy consumption in cellular networks

The objective of this paper is to combine the antenna downtilt selection with the cell size selection in order to reduce the overall radio frequency (RF) transmission power in the homogeneous High-Speed Packet Downlink (HSDPA) cellular radio access network (RAN). The analysis is based on the concept of small cells deployment. The energy consumption ratio (ECR) and the energy reduction gain (ERG) of the cellular RAN are calculated for different antenna tilts when the cell size is being reduced for a given user density and service area. The results have shown that a suitable antenna tilt and the RF power setting can achieve an overall energy reduction of up to 82.56%. Equally, our results demonstrate that a small cell deployment can considerably reduce the overall energy consumption of a cellular network.

Feed-forward neural networks and topographic mappings for exploratory data analysis

A recent novel approach to the visualisation and analysis of datasets, and one which is particularly applicable to those of a high dimension, is discussed in the context of real applications. A feed-forward neural network is utilised to effect a topographic, structure-preserving, dimension-reducing transformation of the data, with an additional facility to incorporate different degrees of associated subjective information. The properties of this transformation are illustrated on synthetic and real datasets, including the 1992 UK Research Assessment Exercise for funding in higher education. The method is compared and contrasted to established techniques for feature extraction, and related to topographic mappings, the Sammon projection and the statistical field of multidimensional scaling.

Secure Transmission in Spectrum Sharing MIMO Networks with Generalized Antenna Selection over Nakagami-m Channels

In this paper, we investigate the secrecy outage performance of spectrum sharing multiple-input multiple-output networks using generalized transmit antenna selection with maximal ratio combining over Nakagami-m channels. In particular, the outdated channel state information is considered at the process of antenna selection due to feedback delay. Considering a practical passive eavesdropper scenario, we derive the exact and asymptotic closed-form expressions of secrecy outage probability, which enable us to evaluate the secrecy performance with high efficiency and present a new design insight into the impact of key parameters on the secrecy performance. In addition, the analytical results demonstrate that the achievable secrecy diversity order is only determined by the parameters of the secondary network, while other parameters related to primary or eavesdropper’s channels have a significantly impact on the secrecy coding gain. 

Neural networks and statistical analysis for classification of corneal videokeratography maps based on Zernike coefficients: a quantitative comparison

PURPOSE: The main goal of this study was to develop and compare two different techniques for classification of specific types of corneal shapes when Zernike coefficients are used as inputs. A feed-forward artificial Neural Network (NN) and discriminant analysis (DA) techniques were used. METHODS: The inputs both for the NN and DA were the first 15 standard Zernike coefficients for 80 previously classified corneal elevation data files from an Eyesys System 2000 Videokeratograph (VK), installed at the Departamento de Oftalmologia of the Escola Paulista de Medicina, São Paulo. The NN had 5 output neurons which were associated with 5 typical corneal shapes: keratoconus, with-the-rule astigmatism, against-the-rule astigmatism, "regular" or "normal" shape and post-PRK. RESULTS: The NN and DA responses were statistically analyzed in terms of precision ([true positive+true negative]/total number of cases). Mean overall results for all cases for the NN and DA techniques were, respectively, 94% and 84.8%. CONCLUSION: Although we used a relatively small database, results obtained in the present study indicate that Zernike polynomials as descriptors of corneal shape may be a reliable parameter as input data for diagnostic automation of VK maps, using either NN or DA.

A beam-forming network for a circular switched-beam phased array antenna

This paper reports on the design and development of a dividing/phasing network for a compact switched-beam array antenna for Land-vehicle mobile satellite communications, The device is formed by a switched radial divider/combiner and 1-bit phase shifters and generates a sufficient number of beams for the proper satellite tracking.