Generative Maximum Entropy Learning for Multiclass Classification

Maximum entropy approach to classification is very well studied in applied statistics and machine learning and almost all the methods that exists in literature are discriminative in nature. In this paper, we introduce a maximum entropy classification method with feature selection for large dimensional data such as text datasets that is generative in nature. To tackle the curse of dimensionality of large data sets, we employ conditional independence assumption (Naive Bayes) and we perform feature selection simultaneously, by enforcing a `maximum discrimination' between estimated class conditional densities. For two class problems, in the proposed method, we use Jeffreys (J) divergence to discriminate the class conditional densities. To extend our method to the multi-class case, we propose a completely new approach by considering a multi-distribution divergence: we replace Jeffreys divergence by Jensen-Shannon (JS) divergence to discriminate conditional densities of multiple classes. In order to reduce computational complexity, we employ a modified Jensen-Shannon divergence (JS(GM)), based on AM-GM inequality. We show that the resulting divergence is a natural generalization of Jeffreys divergence to a multiple distributions case. As far as the theoretical justifications are concerned we show that when one intends to select the best features in a generative maximum entropy approach, maximum discrimination using J-divergence emerges naturally in binary classification. Performance and comparative study of the proposed algorithms have been demonstrated on large dimensional text and gene expression datasets that show our methods scale up very well with large dimensional datasets.

Axial heat conduction effects in natural convection along a vertical cylinder

MANY TRANSPORprTo cesses occur in nature and in industrial applications in which the transfer of heat is governed by the process of natural convection. Natural convection arises in fluids when the temperature changes cause density variations leading to buoyancy forces. An excellent review of natural convection flows has been given by Ede [I]. Recently, Minkowycz and Sparrow [2, 31, Cebeci [4], and Aziz and Na [S] have studied the steady, laminar, incompressible, natural convection flow over a vertical cylinder using a local nonsimilarity method, a finite-difference scheme, and an improved perturbation method, respectively. However, they did not take into account the effect ofaxial heat conduction for small Prandtl number. It is known that the axial heat conductioneffect becomesimportant for low-Prandtl-number fluids such as a liquid metal.

Nonsimilar laminar incompressible boundary layers with vectored mass transfer

The effect of vectored mass transfer on the flow and heat transfer of the steady laminar incompressible nonsimilar boundary layer with viscous dissipation for two-dimensional and axisymmetric porous bodies with pressure gradient has been studied. The partial differential equations governing the flow have been solved numerically using an implicit finite-difference scheme. The computations have been carried out for a cylinder and a sphere. The skin friction is strongly influenced by the vectored mass transfer, and the heat transfer both by the vectored mass transfer and dissipation parameter. It is observed that the vectored suction tends to delay the separation whereas the effect of the vectored injection is just the reverse. Our results agree with those of the local nonsimilarity, difference-differential and asymptotic methods but not with those of the local similarity method.

Compressible magnetohydrodynamic boundary layer swirling nozzle and diffuser flows with a magnetic field

The flow, heat and mass transfer problem for boundary layer swirling flow of a laminar steady compressible electrically conducting gas with variable properties through a conical nozzle and a diffuser with an applied magnetic field has been studied. The partial differential equations governing the flow have been solved numerically using an implicit finite-difference scheme after they have been transformed into dimensionless form using the modified Lees transformation. The results indicate that the skin friction and heat transfer strongly depend on the magnetic field, mass transfer and variation of the density-viscosity product across the boundary layer. However, the effect of the variation of the density-viscosity product is more pronounced in the case of a nozzle than in the case of a diffuser. It has been found that large swirl is required to produce strong effect on the skin friction and heat transfer. Separationless flow along the entire length of the diffuser can be obtained by applying appropriate amount of suction. The results are found to be in good agreement with those of the local nonsimilarity method, but they differ quite significantly from those of the local similarity method.

Compressible Boundary-Layer Swirling Flow in a Nozzle and a Diffuser With Highly Cooled Wall

The steady laminar compressible boundary-layer swirling flow with variable gas properties and mass transfer through a conical nozzle, and a diffuser with a highly cooled wall has been studied. The partial differential equations governing the nonsimilar flow have been transformed to a system of coordinates using modified Lees transformation. The resulting equations are transformed into coordinates having finite ranges by means of a transformation which maps an infinite region into a finite region. The ensuing equations are then solved numerically using an implicit finite-difference scheme. The results indicate that the variation of the density-viscosity product across the boundary layer and mass transfer have strong effect on the skin friction and heat transfer. Separationless flow along the entire length of the diffuser can be obtained by applying suction. The results are found to be in good agreement with those of the local nonsimilarity method but they differ appreciably from those of the local similarity method.

Flow and heat transfer in steady laminar compressible boundary layer swirling flow in a nozzle

The axisymmetric steady laminar compressible boundary layer swirling flow of a gas with variable properties in a nozzle has been investigated. The partial differential equations governing the non-similar flow have been transformed into new co-ordinates having finite ranges by means of a transformation which maps an infinite range into a finite one. The resulting equations have been solved numerically using an implicit finite-difference scheme. The computations have been carried out for compressible swirling flow through a convergent conical nozzle. The results indicate that the swirl exerts a strong influence on the longitudinal skin friction, but its effect on the tangential skin friction and heat transfer is comparatively small. The effect of the variation of the density-viscosity product across the boundary layer is appreciable only at low-wall temperature. The results are in good agreement with those of the local-similarity method for small values of the longitudinal distance.

Boundary layer swirling flow through a conical hydrocyclone

The paper deals with the flow and heat-transfer problem of a steady axisymmetric laminar incompressible boundary layer swirling flow of a fluid through a conical hydrocyclone. The implicit finitedifference scheme is used to solve the partial differential equations governing the flow. The effect of swirl is found to be more pronounced on the longitudinal skin friction than on the tangential skin friction and heat transfer. The skin friction and heat transfer increase with swirl or with longitudinal distance. Swirl also gives rise to velocity overshoot in the longitudinal velocity profiles and the magnitude of the velocity overshoot increases as the swirl parameter increases. The results are found to be in good agreement with those of the local nonsimilarity and momentum integral methods but they differ appreciably from those of the local similarity method except for the longitudinal skin friction which is fairly in good agreement with that of the local similarity method.Die Arbeit beschäftigt sich mit der Strömung und dem Wärmeübergang in einem konischen Zyklon unter der Voraussetzung stationärer, achsensymmetrischer, laminarer, inkompressibler Grenzschichtströmung. Ein implizites Differenzenverfahren wird benutzt, um die partiellen Differentialgleichungen zu lösen. Der Einfluß des Dralls ist besonders ausgeprägt auf die longitudinale Komponente der Oberflächenreibung, weniger dagegen bei der tangentialen Komponente und beim Wärmeübergang. Die Oberflächenreibung und der Wärmeübergang nehmen zu mit dem Drall, sowie mit dem longitudinalen Abstand. Der Drall erzeugt ein Überschießen der Geschwindigkeit in der longitudinalen Abstand. Der Drall erzeugt ein Überschießen der Geschwindigkeit in der Längsrichtung. Die Größe des Überschusses nimmt mit wachsendem Drallparameter zu. Die Resultate stimmen gut mit den Ergebnissen der Theorie der lokalen Nichtähnlichkeit und der Impulsintegralmethode überein. Dagegen weichen sie mit Ausnahme der longitudinalen Komponente der Oberflächenreibung beträchtlich von den Resultaten der Theorie der lokalen Ähnlichkeit ab.

Multivariate nonlinear ensemble prediction of daily chaotic rainfall with climate inputs

The basic characteristic of a chaotic system is its sensitivity to the infinitesimal changes in its initial conditions. A limit to predictability in chaotic system arises mainly due to this sensitivity and also due to the ineffectiveness of the model to reveal the underlying dynamics of the system. In the present study, an attempt is made to quantify these uncertainties involved and thereby improve the predictability by adopting a multivariate nonlinear ensemble prediction. Daily rainfall data of Malaprabha basin, India for the period 1955-2000 is used for the study. It is found to exhibit a low dimensional chaotic nature with the dimension varying from 5 to 7. A multivariate phase space is generated, considering a climate data set of 16 variables. The chaotic nature of each of these variables is confirmed using false nearest neighbor method. The redundancy, if any, of this atmospheric data set is further removed by employing principal component analysis (PCA) method and thereby reducing it to eight principal components (PCs). This multivariate series (rainfall along with eight PCs) is found to exhibit a low dimensional chaotic nature with dimension 10. Nonlinear prediction employing local approximation method is done using univariate series (rainfall alone) and multivariate series for different combinations of embedding dimensions and delay times. The uncertainty in initial conditions is thus addressed by reconstructing the phase space using different combinations of parameters. The ensembles generated from multivariate predictions are found to be better than those from univariate predictions. The uncertainty in predictions is decreased or in other words predictability is increased by adopting multivariate nonlinear ensemble prediction. The restriction on predictability of a chaotic series can thus be altered by quantifying the uncertainty in the initial conditions and also by including other possible variables, which may influence the system. (C) 2011 Elsevier B.V. All rights reserved.

Minimum error solutions of the Boltzmann equation for shock structure

‘Best’ solutions for the shock-structure problem are obtained by solving the Boltzmann equation for a rigid sphere gas by applying minimum error criteria on the Mott-Smith ansatz. The use of two such criteria minimizing respectively the local and total errors, as well as independent computations of the remaining error, establish the high accuracy of the solutions, although it is shown that the Mott-Smith distribution is not an exact solution of the Boltzmann equation even at infinite Mach number. The minimum local error method is found to be particularly simple and efficient. Adopting the present solutions as the standard of comparison, it is found that the widely used v2x-moment solutions can be as much as a third in error, but that results based on Rosen's method provide good approximations. Finally, it is shown that if the Maxwell mean free path on the hot side of the shock is chosen as the scaling length, the value of the density-slope shock thickness is relatively insensitive to the intermolecular potential. A comparison is made on this basis of present results with experiment, and very satisfactory quantitative agreement is obtained.

Understanding the Magnetization Reversal in Six-Fold Anisotropic Hexagonal Networks

Isolated magnetic nanowires have been studied extensively and the magnetization reversal mechanism is well understood in these systems. But when these nanowires are joined together in different architectures, they behave differently and can give novel properties. Using this approach, one can engineer the network architectures to get artificial anisotropy. Here, we report six-fold anisotropy by joining the magnetic nanowires into hexagonal network. For this study, we also benchmark the widely used micromagnetic packages: OOMMF, Nmag, and LLG-simulator. Further, we propose a local hysteresis method by post processing the spatial magnetization information. With this approach we obtained the hysteresis of nanowires to understand the six-fold anisotropy and the reversal mechanism within the hexagonal networks.

A new method of producing local enhancement of buoyancy in liquid flows

We describe here a novel method of generating large volumetric heating in a liquid. The method uses the principle of ohmic heating of the liquid, rendered electrically conducting by suitable additives if necessary. Electrolysis is prevented by the use of high frequency alternating voltage and chemically treated electrodes. The technique is demonstrated by producing substantial heating in an initially neutral jet of water. Simple flow visualisation studies, made by adding dye to the jet, show marked changes in the growth and development of the jet with heat addition.

Site Characterization of Bangalore using Geophysical Method

A number of geophysical methods have been proposed for near-surface site characterization and measurement of shear wave velocity by using a great variety of testing configurations, processing techniques,and inversion algorithms. In particular, two widely-used techniques are SASW (Spectral Analysis of SurfaceWaves) and MASW (Multichannel Analysis of SurfaceWaves). MASW is increasingly being applied to earthquake geotechnical engineering for the local site characterization, microzonation and site response studies.A MASW is a geophysical method, which generates a shear-wave velocity (Vs) profile (i.e., Vs versus depth)by analyzing Raleigh-type surface waves on a multichannel record. MASW system consisting of 24 channels Geode seismograph with 24 geophones of 4.5 Hz frequency have been used in this investigation. For the site characterization program, the MASW field experiments consisting of 58 one-dimensional shear wave velocity tests and 20 two-dimensional shear wave tests have been carried out. The survey points have been selected in such a way that the results supposedly represent the whole metropolitan Bangalore having an area of 220 km2.The average shear wave velocity of Bangalore soils have been evaluated for depths of 5m, 10m, 15m, 20m, 25m and 30 m. The subsoil site classification has been made for seismic local site effect evaluation based on average shear wave velocity of 30m depth (Vs30) of sites using National Earthquake Hazards Reduction Program (NEHRP) and International Building Code (IBC) classification. Soil average shearwave velocity estimated based on overburden thickness from the borehole information is also presented. Mapping clearly indicates that the depth of soil obtained from MASW is closely matching with the soil layers in bore logs. Among total 55 locations of MASW survey carried out, 34 locations were very close to the SPT borehole locations and these are used to generate correlation between Vs and corrected “N” values. The SPT field “N” values are corrected by applying the NEHRP recommended corrections.

Method for Seismic Microzonation with Geotechnical Aspects

This study presents an overview of seismic microzonation and existing methodologies with a newly proposed methodology covering all aspects. Earlier seismic microzonation methods focused on parameters that affect the structure or foundation related problems. But seismic microzonation has generally been recognized as an important component of urban planning and disaster management. So seismic microzonation should evaluate all possible hazards due to earthquake and represent the same by spatial distribution. This paper presents a new methodology for seismic microzonation which has been generated based on location of study area and possible associated hazards. This new method consists of seven important steps with defined output for each step and these steps are linked with each other. Addressing one step and respective result may not be seismic microzonation, which is practiced widely. This paper also presents importance of geotechnical aspects in seismic microzonation and how geotechnical aspects affect the final map. For the case study, seismic hazard values at rock level are estimated considering the seismotectonic parameters of the region using deterministic and probabilistic seismic hazard analysis. Surface level hazard values are estimated considering site specific study and local site effects based on site classification/characterization. The liquefaction hazard is estimated using standard penetration test data. These hazard parameters are integrated in Geographical Information System (GIS) using Analytic Hierarchy Process (AHP) and used to estimate hazard index. Hazard index is arrived by following a multi-criteria evaluation technique - AHP, in which each theme and features have been assigned weights and then ranked respectively according to a consensus opinion about their relative significance to the seismic hazard. The hazard values are integrated through spatial union to obtain the deterministic microzonation map and probabilistic microzonation map for a specific return period. Seismological parameters are widely used for microzonation rather than geotechnical parameters. But studies show that the hazard index values are based on site specific geotechnical parameters.

Gene expression programming-fuzzy logic method for crop type classification

Crop type classification using remote sensing data plays a vital role in planning cultivation activities and for optimal usage of the available fertile land. Thus a reliable and precise classification of agricultural crops can help improve agricultural productivity. Hence in this paper a gene expression programming based fuzzy logic approach for multiclass crop classification using Multispectral satellite image is proposed. The purpose of this work is to utilize the optimization capabilities of GEP for tuning the fuzzy membership functions. The capabilities of GEP as a classifier is also studied. The proposed method is compared to Bayesian and Maximum likelihood classifier in terms of performance evaluation. From the results we can conclude that the proposed method is effective for classification.

Efficient method of moving shadow detection and vehicle classification

Moving shadow detection and removal from the extracted foreground regions of video frames, aim to limit the risk of misconsideration of moving shadows as a part of moving objects. This operation thus enhances the rate of accuracy in detection and classification of moving objects. With a similar reasoning, the present paper proposes an efficient method for the discrimination of moving object and moving shadow regions in a video sequence, with no human intervention. Also, it requires less computational burden and works effectively under dynamic traffic road conditions on highways (with and without marking lines), street ways (with and without marking lines). Further, we have used scale-invariant feature transform-based features for the classification of moving vehicles (with and without shadow regions), which enhances the effectiveness of the proposed method. The potentiality of the method is tested with various data sets collected from different road traffic scenarios, and its superiority is compared with the existing methods. (C) 2013 Elsevier GmbH. All rights reserved.