Abstract: In this paper, we propose the variational EM inference
algorithm for the multi-class Gaussian process classification model
that can be used in the field of human behavior recognition. This
algorithm can drive simultaneously both a posterior distribution of a
latent function and estimators of hyper-parameters in a Gaussian
process classification model with multiclass. Our algorithm is based
on the Laplace approximation (LA) technique and variational EM
framework. This is performed in two steps: called expectation and
maximization steps. First, in the expectation step, using the Bayesian
formula and LA technique, we derive approximately the posterior
distribution of the latent function indicating the possibility that each
observation belongs to a certain class in the Gaussian process
classification model. Second, in the maximization step, using a derived
posterior distribution of latent function, we compute the maximum
likelihood estimator for hyper-parameters of a covariance matrix
necessary to define prior distribution for latent function. These two
steps iteratively repeat until a convergence condition satisfies.
Moreover, we apply the proposed algorithm with human action
classification problem using a public database, namely, the KTH
human action data set. Experimental results reveal that the proposed
algorithm shows good performance on this data set.
Abstract: The Com-Poisson (CMP) model is one of the most
popular discrete generalized linear models (GLMS) that handles
both equi-, over- and under-dispersed data. In longitudinal context,
an integer-valued autoregressive (INAR(1)) process that incorporates
covariate specification has been developed to model longitudinal
CMP counts. However, the joint likelihood CMP function is
difficult to specify and thus restricts the likelihood-based estimating
methodology. The joint generalized quasi-likelihood approach
(GQL-I) was instead considered but is rather computationally
intensive and may not even estimate the regression effects due
to a complex and frequently ill-conditioned covariance structure.
This paper proposes a new GQL approach for estimating the
regression parameters (GQL-III) that is based on a single score vector
representation. The performance of GQL-III is compared with GQL-I
and separate marginal GQLs (GQL-II) through some simulation
experiments and is proved to yield equally efficient estimates as
GQL-I and is far more computationally stable.
Abstract: The present study aimed to determine the
effectiveness of Metaphor therapy on depression among female
students. The sample included 60 female students with depression
symptoms selected by simple sampling and randomly divided into
two equal groups (experimental and control groups). Beck
Depression Inventory was used to measure the variables. This was an
experimental study with a pre-test/post-test design with control
group. Eight metaphor therapy sessions were held for the
experimental group. A post-test was administered to both groups.
Data were analyzed using multivariate analysis of covariance
(MANCOVA). Results showed that the Metaphor therapy decreased
depression in the experimental group compared to the control group.
Abstract: The purpose of the study was to find out the effects of
Aquatic and Land plyometric training on selected physical variables
in intercollegiate male handball players. To achieve this purpose of
the study, forty five handball players of Sardar Vallabhbhai National
Institute of Technology, Surat, Gujarat were selected as players at
random and their age ranged between 18 to 21 years. The selected
players were divided into three equal groups of fifteen players each.
Group I underwent Aquatic plyometric training, Group II underwent
Land plyometric training and Group III Control group for three days
per week for twelve weeks. Control Group did not participate in any
special training programme apart from their regular activities as per
their curriculum. The following physical fitness variables namely
speed; leg explosive power and agility were selected as dependent
variables. All the players of three groups were tested on selected
dependent variables prior to and immediately after the training
programme. The analysis of covariance was used to analyze the
significant difference, if any among the groups. Since, three groups
were compared, whenever the obtained ‘F’ ratio for adjusted posttest
was found to be significant, the Scheffe’s test to find out the paired
mean differences, if any. The 0.05 level of confidence was fixed as
the level of significance to test the ‘F’ ratio obtained by the analysis
of covariance, which was considered as an appropriate. The result of
the study indicates due to Aquatic and Land plyometric training on
speed, explosive power, and agility has been improved significantly.
Abstract: In statistics parameter theory, usually the
parameter estimations have two kinds, one is the least-square
estimation (LSE), and the other is the best linear unbiased
estimation (BLUE). Due to the determining theorem of
minimum variance unbiased estimator (MVUE), the parameter
estimation of BLUE in linear model is most ideal. But since
the calculations are complicated or the covariance is not
given, people are hardly to get the solution. Therefore, people
prefer to use LSE rather than BLUE. And this substitution
will take some losses. To quantize the losses, many scholars
have presented many kinds of different relative efficiencies in
different views. For the linear weighted regression model, this
paper discusses the relative efficiencies of LSE of β to BLUE
of β. It also defines two new relative efficiencies and gives
their lower bounds.
Abstract: The problems arising from unbalanced data sets
generally appear in real world applications. Due to unequal class
distribution, many researchers have found that the performance of
existing classifiers tends to be biased towards the majority class. The
k-nearest neighbors’ nonparametric discriminant analysis is a method
that was proposed for classifying unbalanced classes with good
performance. In this study, the methods of discriminant analysis are
of interest in investigating misclassification error rates for classimbalanced
data of three diabetes risk groups. The purpose of this
study was to compare the classification performance between
parametric discriminant analysis and nonparametric discriminant
analysis in a three-class classification of class-imbalanced data of
diabetes risk groups. Data from a project maintaining healthy
conditions for 599 employees of a government hospital in Bangkok
were obtained for the classification problem. The employees were
divided into three diabetes risk groups: non-risk (90%), risk (5%),
and diabetic (5%). The original data including the variables of
diabetes risk group, age, gender, blood glucose, and BMI were
analyzed and bootstrapped for 50 and 100 samples, 599 observations
per sample, for additional estimation of the misclassification error
rate. Each data set was explored for the departure of multivariate
normality and the equality of covariance matrices of the three risk
groups. Both the original data and the bootstrap samples showed nonnormality
and unequal covariance matrices. The parametric linear
discriminant function, quadratic discriminant function, and the
nonparametric k-nearest neighbors’ discriminant function were
performed over 50 and 100 bootstrap samples and applied to the
original data. Searching the optimal classification rule, the choices of
prior probabilities were set up for both equal proportions (0.33: 0.33:
0.33) and unequal proportions of (0.90:0.05:0.05), (0.80: 0.10: 0.10)
and (0.70, 0.15, 0.15). The results from 50 and 100 bootstrap samples
indicated that the k-nearest neighbors approach when k=3 or k=4 and
the defined prior probabilities of non-risk: risk: diabetic as 0.90:
0.05:0.05 or 0.80:0.10:0.10 gave the smallest error rate of
misclassification. The k-nearest neighbors approach would be
suggested for classifying a three-class-imbalanced data of diabetes
risk groups.
Abstract: Subspace channel estimation methods have been
studied widely, where the subspace of the covariance matrix is
decomposed to separate the signal subspace from noise subspace. The
decomposition is normally done by using either the eigenvalue
decomposition (EVD) or the singular value decomposition (SVD) of
the auto-correlation matrix (ACM). However, the subspace
decomposition process is computationally expensive. This paper
considers the estimation of the multipath slow frequency hopping
(FH) channel using noise space based method. In particular, an
efficient method is proposed to estimate the multipath time delays by
applying multiple signal classification (MUSIC) algorithm which is
based on the null space extracted by the rank revealing LU (RRLU)
factorization. As a result, precise information is provided by the
RRLU about the numerical null space and the rank, (i.e., important
tool in linear algebra). The simulation results demonstrate the
effectiveness of the proposed novel method by approximately
decreasing the computational complexity to the half as compared
with RRQR methods keeping the same performance.
Abstract: This study is purposed to develop an efficient fault
detection method for Global Navigation Satellite Systems (GNSS)
applications based on adaptive noise covariance estimation. Due to the
dependence on radio frequency signals, GNSS measurements are
dominated by systematic errors in receiver’s operating environment.
In the proposed method, the pseudorange and carrier-phase
measurement noise covariances are obtained at time propagations and
measurement updates in process of Carrier-Smoothed Code (CSC)
filtering, respectively. The test statistics for fault detection are
generated by the estimated measurement noise covariances. To
evaluate the fault detection capability, intentional faults were added to
the filed-collected measurements. The experiment result shows that
the proposed method is efficient in detecting unhealthy measurements
and improves GNSS positioning accuracy against fault occurrences.
Abstract: In this paper two approaches to joint signal detection,
time of arrival (ToA) and angle of arrival (AoA) estimation in
multi-element antenna array are investigated. Two scenarios were
considered: first one, when the waveform of the useful signal
is known a priori and, second one, when the waveform of the
desired signal is unknown. For first scenario, the antenna array
signal processing based on multi-element matched filtering (MF)
with the following non-coherent detection scheme and maximum
likelihood (ML) parameter estimation blocks is exploited. For second
scenario, the signal processing based on the antenna array elements
covariance matrix estimation with the following eigenvector analysis
and ML parameter estimation blocks is applied. The performance
characteristics of both signal processing schemes are thoroughly
investigated and compared for different useful signals and noise
parameters.
Abstract: This paper proposes a GLMM with spatial and
temporal effects for malaria data in Thailand. A Bayesian method is
used for parameter estimation via Gibbs sampling MCMC. A
conditional autoregressive (CAR) model is assumed to present the
spatial effects. The temporal correlation is presented through the
covariance matrix of the random effects. The malaria quarterly data
have been extracted from the Bureau of Epidemiology, Ministry of
Public Health of Thailand. The factors considered are rainfall and
temperature. The result shows that rainfall and temperature are
positively related to the malaria morbidity rate. The posterior means
of the estimated morbidity rates are used to construct the malaria
maps. The top 5 highest morbidity rates (per 100,000 population) are
in Trat (Q3, 111.70), Chiang Mai (Q3, 104.70), Narathiwat (Q4,
97.69), Chiang Mai (Q2, 88.51), and Chanthaburi (Q3, 86.82).
According to the DIC criterion, the proposed model has a better
performance than the GLMM with spatial effects but without
temporal terms.
Abstract: This paper presents a nonparametric identification of
continuous-time nonlinear systems by using a Gaussian process
(GP) model. The GP prior model is trained by artificial bee colony
algorithm. The nonlinear function of the objective system is estimated
as the predictive mean function of the GP, and the confidence
measure of the estimated nonlinear function is given by the predictive
covariance of the GP. The proposed identification method is applied
to modeling of a simplified electric power system. Simulation results
are shown to demonstrate the effectiveness of the proposed method.
Abstract: Nanotechnology is widely applied in various areas so professionals in the related fields have to know more than nano knowledge. In the study, we focus on adopting ICT-assisted PBL in college general education to foster professionals who possess multiple abilities. The research adopted a pretest and posttest quasi-experimental design. The control group received traditional instruction, and the experimental group received ICT-assisted PBL instruction. Descriptive statistics will be used to describe the means, standard deviations, and adjusted means for the tests between the two groups. Next, analysis of covariance (ANCOVA) will be used to compare the final results of the two research groups after 6 weeks of instruction. Statistics gathered in the end of the research can be used to make contrasts. Therefore, we will see how different teaching strategies can improve students’ understanding about nanotechnology and learning skills.
Abstract: The two common approaches to Structural Equation Modeling (SEM) are the Covariance-Based SEM (CB-SEM) and Partial Least Squares SEM (PLS-SEM). There is much debate on the performance of CB-SEM and PLS-SEM for small sample size and when distributions are nonnormal. This study evaluates the performance of CB-SEM and PLS-SEM under normality and nonnormality conditions via a simulation. Monte Carlo Simulation in R programming language was employed to generate data based on the theoretical model with one endogenous and four exogenous variables. Each latent variable has three indicators. For normal distributions, CB-SEM estimates were found to be inaccurate for small sample size while PLS-SEM could produce the path estimates. Meanwhile, for a larger sample size, CB-SEM estimates have lower variability compared to PLS-SEM. Under nonnormality, CB-SEM path estimates were inaccurate for small sample size. However, CB-SEM estimates are more accurate than those of PLS-SEM for sample size of 50 and above. The PLS-SEM estimates are not accurate unless sample size is very large.
Abstract: Antenna arrays are widely used in modern radio systems in sonar and communications. The solving of the detection problems of a useful signal on the background of noise is based on the GLRT method. There is a large number of problem which depends on the known a priori information. In this work, in contrast to the majority of already solved problems, it is used only difference spatial properties of the signal and noise for detection. We are analyzing the influence of the degree of non-coherence of signal and noise unhomogeneity on the performance characteristics of different GLRT statistics. The description of the signal and noise is carried out by means of the spatial covariance matrices C in the cases of different number of known information. The partially coherent signalis is simulated as a plane wave with a random angle of incidence of the wave concerning a normal. Background noise is simulated as random process with uniform distribution function in each element. The results of investigation of degradation of performance characteristics for different cases are represented in this work.
Abstract: The traditional second order statistics approach of using only the hermitian covariance for non circular signals, does not take advantage of the information contained in the complementary covariance of these signals. Radar systems often use non circular signals such as Binary Phase Shift Keying (BPSK) signals. Their noncicular property can be exploited together with the dual centrosymmetry of the bistatic MIMO radar system to improve angle estimation performance. We construct an augmented matrix from the received data vectors using both the positive definite hermitian covariance matrix and the complementary covariance matrix. The Unitary ESPRIT technique is then applied to the signal subspace of the augmented covariance matrix for automatically paired Direction-of-arrival (DOA) and Direction-of-Departure (DOD) angle estimates. The number of targets that can be detected is twice that obtainable with the conventional ESPRIT approach. Simulation results show the effectiveness of this method in terms of increase in resolution and the number of targets that can be detected.
Abstract: Based on the sources- smoothed rank profile (SRP) and modified minimum description length (MMDL) principle, a method for estimation of the source coherency structure (SCS) and the number of wideband sources is proposed in this paper. Instead of focusing, we first use a spatial smoothing technique to pre-process the array covariance matrix of each frequency for de-correlating the sources and then use smoothed rank profile to determine the SCS and the number of wideband sources. We demonstrate the availability of the method by numerical simulations.
Abstract: When the failure function is monotone, some monotonic reliability methods are used to gratefully simplify and facilitate the reliability computations. However, these methods often work in a transformed iso-probabilistic space. To this end, a monotonic simulator or transformation is needed in order that the transformed failure function is still monotone. This note proves at first that the output distribution of failure function is invariant under the transformation. And then it presents some conditions under which the transformed function is still monotone in the newly obtained space. These concern the copulas and the dependence concepts. In many engineering applications, the Gaussian copulas are often used to approximate the real word copulas while the available information on the random variables is limited to the set of marginal distributions and the covariances. So this note catches an importance on the conditional monotonicity of the often used transformation from an independent random vector into a dependent random vector with Gaussian copulas.
Abstract: In this paper we propose a novel method for human
face segmentation using the elliptical structure of the human head. It
makes use of the information present in the edge map of the image.
In this approach we use the fact that the eigenvalues of covariance
matrix represent the elliptical structure. The large and small
eigenvalues of covariance matrix are associated with major and
minor axial lengths of an ellipse. The other elliptical parameters are
used to identify the centre and orientation of the face. Since an
Elliptical Hough Transform requires 5D Hough Space, the Circular
Hough Transform (CHT) is used to evaluate the elliptical parameters.
Sparse matrix technique is used to perform CHT, as it squeeze zero
elements, and have only a small number of non-zero elements,
thereby having an advantage of less storage space and computational
time. Neighborhood suppression scheme is used to identify the valid
Hough peaks. The accurate position of the circumference pixels for
occluded and distorted ellipses is identified using Bresenham-s
Raster Scan Algorithm which uses the geometrical symmetry
properties. This method does not require the evaluation of tangents
for curvature contours, which are very sensitive to noise. The method
has been evaluated on several images with different face orientations.
Abstract: A clustering is process to identify a homogeneous
groups of object called as cluster. Clustering is one interesting topic
on data mining. A group or class behaves similarly characteristics.
This paper discusses a robust clustering process for data images with
two reduction dimension approaches; i.e. the two dimensional
principal component analysis (2DPCA) and principal component
analysis (PCA). A standard approach to overcome this problem is
dimension reduction, which transforms a high-dimensional data into
a lower-dimensional space with limited loss of information. One of
the most common forms of dimensionality reduction is the principal
components analysis (PCA). The 2DPCA is often called a variant of
principal component (PCA), the image matrices were directly treated
as 2D matrices; they do not need to be transformed into a vector so
that the covariance matrix of image can be constructed directly using
the original image matrices. The decomposed classical covariance
matrix is very sensitive to outlying observations. The objective of
paper is to compare the performance of robust minimizing vector
variance (MVV) in the two dimensional projection PCA (2DPCA)
and the PCA for clustering on an arbitrary data image when outliers
are hiden in the data set. The simulation aspects of robustness and
the illustration of clustering images are discussed in the end of
paper
Abstract: The purpose of this study was to investigate the effects of computer–based instructional designs, namely modality and redundancy principles on the attitude and learning of music theory among primary pupils of different Music Intelligence levels. The lesson of music theory was developed in three different modes, audio and image (AI), text with image (TI) and audio with image and text (AIT). The independent variables were the three modes of courseware. The moderator variable was music intelligence. The dependent variables were the post test score. ANOVA was used to determine the significant differences of the pretest scores among the three groups. Analyses of covariance (ANCOVA) and Post hoc were carried out to examine the main effects as well as the interaction effects of the independent variables on the dependent variables. High music intelligence pupils performed significantly better than low music intelligence pupils in all the three treatment modes. The AI mode was found to help pupils with low music intelligence significantly more than the TI and AIT modes.