IRAQI JOURNAL OF STATISTICAL SCIENCES

In this paper we have proposed a modified hybrid conjugate direction algorithm which combined a well-known CG-method which was based on a non-quadratic model and a well-known VM-method which based on the quadratic model.
The new modified algorithm was treated both theoretically and numerically and proved to be stable and its convergence was super-linear and it uses an exact line search.
Our numerical results indicate that the modified hybrid method performs well compared to the two well-known used method.

This paper is devoted to estimate the error resulting from approximating crack, i.e. approximating the support of the jump of crack in mobile's wire. This estimation is important to identify the interfaces crack which have been completed by using the derivation of Reciprocity Gap of this model. This derivation is based on the principle of Galerkin finite element method.

In this paper, we have investigated a new multiplier for the Lagrange interpolation by modifying the initial value of the multiplier in order to reduce the errors and avoid the use of arbitrary values for the initial . The new procedure improves the rate of convergence of the Lagrange interpolation. Numerical results indicate that the new approach yields very effective numerical results depending on the number of iterations ; number of gradient and the number of function calls.

This research deals with building a stochastic model for the sequence of music notations by considering them as a Markov chain, after explaining some theoretical sides related to music science as well as the basics of Markov chain, specially that related to the concept of order. An application on a well-known Islamic song is considered. The parameters of the Markovian model are estimated as well as the order.

In this paper, we present a modified algorithm with curve search rule for unconstrained minimization problems. At each iteration, the next iterative point is determined by means of a curve search rule. That the search direction and the step-size are particularly determined simultaneously at each iteration of the new algorithm.

In This paper we consider two dimensional flow to a viscous symmetric thin liquid film flow on solid surface with no inertia force . Navier – stokes equation has been used to find the equation that governs this flow we solve these equations analytically.

A model of heat transfer by natural convection of dissipative fluid in a channel of porous walls has been discussed, the solution of governing partial differential equations was obtained using Alternating Direction Implicit method. The unsteady state as well as steady state solutions are founds simultaneously during the successive iteration of (ADI ) for the first time .

Following Mathiew, Cabrera-Rodriguez and Cabera-Mohammed we define particular totally multiplicatively prime algebra. Using these algebras we improved the result due to Cabera-Mohammed which stated that, if A is a real totally multiplicatively prime algebra, than the central closure of A is a complex totally multiplicatively prime algebra.
Also we explain the relation between particular totally multiplicatively prime algebra with prior algebra specifically totally multiplicatively prime algebras ultraprime algebras and totally prime algebras.

Numerical Solution of Korteweg-de Vries-Burger (KdV-B) equation is presented using two finite difference methods ,the explicit scheme and Crank-Nicholson scheme. The accuracy of computed solutions is examined by comparison with analytical solution using example , and it has been found that the explicit scheme is simpler while the Crank-Nicholson is more accurate and has faster convergent . Also , the stability analysis of the two methods by using Fourier (Von Neumann) method has been done and the result was showed that the explicit scheme is stable under the condition and the Crank-Nicholson is unconditionally stable.

The paper considers stochastic convergence of certain means used to obtain the between sum of squares in analysis of variance. These limiting random variables are used to obtain a nonnegative estimator of the between component of variance. The usual ANOVA estimator of the within component of variance is considered. A nonnegative estimator of heritability is given. Asymptotic tests are derived also. Finally, the paper extends the linear model to allow the number of observations in each cell to be random.

In this article we make a comparison between three different nonparametric density estimators .The first estimator is called fixed kernel which uses a fixed bandwidth.
The second estimator is called variable kernel, which uses variable bandwidth, then the third estimator, refers to a new version of nonparametric density estimator (this estimator is a hybrid of the above estimators.
This comparison depends on using different kernels, two of these functions are proposed in this article .To make this comparison suitable, we use standardization techniques, which are called canonical kernel.
Simulation is used to find kernel function based on the standardization techniques (canonical kernel).

This research deals with the non-homogeneous Poisson process in a suggested method through mixing two methods of estimation in one method to arrive at the estimate of the rate of occurrence without using common probability distribution and estimating their parameters. A non-parametric method has been applied in a suggested method of estimation.
The practical application includes studying the period of daily rainfall in Mosul Forecasting Station in Neneva. The rate of daily rainfall has been estimated in the static method suggested in the research. The result arrived at in this research has been compared with other results shown by other two estimators. The efficiency of this suggested method becomes clear in estimating the rate of occurrence of events in non-homogenous Poisson process through this comparison.
Keywords: Non-Homogenous Poisson Process, Kernel Estimation, Piecewise Linear Estimation, Rainfall

The straight lines exist richly in civil objects which are usually man made like buildings, read,…ect. Depending on this fact search strategy was constructed. New method has been used to detecte straight line, it is Baron's Method (BM), after using and checking the results we adopted the (BM) method in this search. Some development are performed on the (BM) method and we called it Developed Baron method (DBM). Hybrid has been done between (DBM) and genetic algorithm (GA) to get efficient results (execution and precision), this method is being called Genetic Developed Baron Method(GDBM).The two methods have been applied. The accuracy and execution time for each method is calculated. The experiments show that the proposed hybrid method in this search is relatively fast and it achieves high performance, it produces (90%) recognition rate.

In this research, A proposed method was used to shrink the wavelet, in order to deals with the problem of noise (or contamination) that may encounter the time series data, and trying to find the best estimation for the parameters of the autoregressive model of order (p) by getting the least possible value (best value) of the estimated statistical criteria's values for the proposed models (of less contamination) as the final prediction error, loss function and The Mean Absolute Prediction Error, and comparing them with the estimated statistical criteria's for the autoregressive model from order (p) that estimated by the classical method.
In the application side of the research, the analysis of the time series observations of the quantity of the annual rainfall in Erbil city for the period (1992-2007) was discussed depending on the use of MATLAB software, in addition to write code using MATLAB language.

In this paper, we present a new computational algorithm which can be used to study the sensitivity analysis of GPP and SPP of degree of difficulty greater than zero. To study the sensitivity analysis of GPP (Dinkal & Kohgenberger [2]), one has to use a subroutine to solve a system of linear equation to get the values of optimal primal variables. Our proposed algorithm is thus more effective in that no new subroutines (such as one for computing the inverse matrices are needed). The generalized algorithm which was presented by Mohan & Al-Bayati to study the sensitivity analysis of SPP [1], is same as our proposed algorithm except that in evaluating the new values of the dual optimal variables. The numerical evidence confirm that the new proposed algorithm is very effective.

In this research the estimation of the regression parameters was clarified using least squares and partial least squares compared with the normal method in terms of its ability to be free from the problem of Multicollinearity between the predictive variables, the first method is the possibility of a regression analysis of several response variables with a number of predictive variables at the same time, was the application of both methods on the data production of cement as described in the manufacture Material of cement, either predictive variables for the response variables by cement and clinker, dust and solid waste.

In this paper, we present extension forms of Dai, Yuan (DY) and Fletcher, Reveres (FR) CG algorithms. Our modifications and based on introducing a non-quadratic model (sigmoid function model). These modified algorithms are implemented with Wolfe conditions, where initial step size in each iteration is taken as and the global convergence of the modified DY algorithm is investigated. These modified algorithms are tested on some standard test functions and compared with the original DY and FR algorithms showing considerable improvements over all these comparisons.

This research deals with the estimation of linear random two-way classification model with interaction, containing four parameters. These are: variance of effect A, variance of effect variance of interaction AB and variance of random error. These parameters are estimated through using Bayes Quadratic Unbiased Estimator (BAQUE).
The prior information obtained by using variance analysis technique to represent prior estimates of these parameters. Then, the prior distribution is considered as a uniform distribution.
BAQUE approach is applied to real data obtained from Mosul University/ College of Agriculture and Forestry / Department of Crops, and these data represent the development of planting yellow corn that the development has three factors which are the corn type, the quantity of nitrogen fertilizer and the interaction between them. Then a random sample was taken from these data to get the random model. The results of estimates that have been obtained are very encouraging.

The Broyden class of Quasi-Newton updates is an approximation for the inverse Hessian matrix, are transformed to the standard BFGS update, which makes it possible to generalize the well-known t-BFGS formula. One of the variants, the simpler of them, is given in this study and does not require any additional matrix storage by vector multiplications. Experimental results indicate that the new proposed algorithm was more efficient than the standard BFGS algorithm.

Spatial Stochastic processes are often divided into two types that are stationary spatial stochastic processes and other nonstationary.
In most practical applications the processes are nonstationary .Prediction of these processes are performed by universal Kriging technique which supposes that the processes have a trend with fixed linear or nonlinear equation.
In this work we derive the universal Kriging equations in matrixes form which can easily be implemented by computer. Also here we identify the variogram model and trend model.
The application is given in example that worked out on real data. The predicted results are very near to the real data with error of prediction is given

This paper introduces an automatic human gender classification (male or female ) depending on ultrasound images using artificial neural network to classify the gender . This system consists of three units : preprocessing the ultrasound images of human gender (noise removed , normalization and segmentation ), Feature extraction and gender classification. After preprocessing the ultrasound images then extracting the features by Kernel principal component analysis (kernel PCA) after that, linear neural network used to a classify training and testing these gender images . The system produces promising results for gender classification .

This paper describes the design and development of a prototype system for the automatic identification of an individual based on the fusion of hand geometry with backhand patterns. Information fusion at the feature extraction and at the confidence level, where the matching scores reported by Bayesian backpropagation neural network, is discussed.
The system was tested with the template files. The test performance, False Acceptance Rate (FAR) = 10% and False Rejection Rate (FRR) = 0%, suggests that the system can be used in medium/high security large buildings environments.

The power and activity of statistical models is often improved after doing a suitable transformation, the efficacy is reflected also on analysis and forecasting using time series models. The mathematical modification of the data raises problems for the interpretation of models results and usefulness of forecasts based on transformed data, so in this paper we deal with methods for retransformed forecasts back into original scale. The application side of this paper will be in daily average data of electric power consumption in Nineveh governorate.

In this research we find alternative ratio estimators to the suggested estimators for [Sesodia and Dwivedi] , [Singh and Kakran] and [Upadhyaya and Singh] , in simple random sampling by using (K) factor. We obtain mean square error (MSE) equation for all proposed estimators and find the theoretical conditions for (K) factor , which makes our proposed estimator more precise than the traditional and suggested estimators .


Keywords: Simple Random Sampling ; Ratio Estimator ; (K) Factor ; Mean Square Errors

The mechanism of missing data requires knowing the reason behind missing them because different mechanism easy in which data is missed give rise to different specifications of the available sample. The data on which this study depends has been missed completely at random (MCAR) and changed into complete one. It can also be possible to find the magnitude of unbias after changing the data into complete one. Then observing the problem of multi linear variance which can be detected by Variance Inflation Factor (VIF) gauge. The estimation of data by the principal component, method follows that. Finally the study applies this data on the variants of oil products and the consumption of the refined products in certain regions in the world (1000 barrels a day from 1970 to 2004).

This paper specifies econometric model for environmental pollution such as solid waste and gas emissions of Kirkuk cement factory for the period (1984-2006) the variables used in the production process, which is electric power, black oil, clay and limestone. We used ridge regression and Generalized Maximum Entropy(GME) to solve the multicollinearity problem,because this problem is appeared between the variables of the model. The multicollinearity diagnostic is done by variance inflation factor,condition index and variance proportions. From the analysis, we obtained the GME as the best estimation with respect to standard error of the parameters and logic. The limestone has the largest influence in solid pollution and the black oil and electric power are the variables which affects gas emissions. The (SAS.9) is used in the statistical analysis.

In this paper, a lower bound for the optimal makespan value have been improved (Hybrid lower bound) which can be used to evaluate the performance algorithm for multi-processor task scheduling problems in
multi-stage FSMP(flow shop with multi-processor).

So, the values of hybrid lower bounds can be used as an optimal solution when the problems are small size.

After applying the hybrid bounds, the result has shown that the bounds are efficient in estimating the optimal makespen value.

In addition, the high competence of this program in calculating the lower bound and execution speed are proofed

This Research comprises building mathematical model for virtual wall design by using two safety constraints : overturning safety factor and sliding safety factor, also reliability's constraints for each factor and using decomposition technique especially bilevel decomposition – relaxation algorithm to obtain the optimal dimensions for the wall, A program has been achieved for designing a wall with the constraints mentioned above via programming relaxation algorithm steps by using Matlab and using optimization toolbox functions in order to find a solution for master and subproblems. Tthis program shows high efficiency and easiness in obtaining results.

This Research aims to study time series and the ability to use transformations, and the use of Exponential Smoothing for prediction formulation with application.

The cross over design with a missing period was split into a set of Youden squares. A direct and indirect methods were applied before and after splitting and a comparation between the results was established. The data have been generated according to an experiment conducted in cross over design which was applied in the college of Agriculture and Forestry, University of Mosul.
The analysis of a set of Youden squares showed a better results through the direct method according to comparative measurements: rising the F value, decreasing both the significant level ( ) and the mean squares error (MSe). Also the analysis of a set of Youden squares showed better results than cross over design analysis through both analytical methods according to the comparative measurements.

In this research, Markov Random Fields Models have been used in images processing, included algorithm suggested for image segmentation tha depends on the triple mixture normal distribution.
Two ways had been demonstrated for this segmentation. The first one K-Means unsupervised technique and the other used Maximum A Posterior, and to be generalized to fourfold and fivefold mixture normal distribution, then to n – mixture normal distribution.
Matlab Package version (7.0) was used for writing the programs of the algorithms.

Queuing theory and other of discrete and continuous event systems have theory aesthetic but the more assumptions which indepent these theories are not accurate as compared with real situation, also the queue problem is one of the most scientific problems that require simulation to analysi it. Because of that the aim of this research is to design event scheduling for simulation of discrete event systems and according to simulation time with application on simple and multiple queuing using MATLAB program (V.7.0)

This paper investigates the analysis of scheduling algorithm of independed processing like the Longest Processing Time Scheduling algorithm(LPT) and Shortest Processing Time Scheduling algorithm(SPT) and walking a comparison between them with a proposed algorithm depending on the measuring length scheduling of each of algorithms to get the ideal solution through applying some examples.

Many sports activities which depend on sport movement and its relation with features ,that may determined through various ways ,for studying speciality of human movement and the physical appearance. And from these phenomena , physical appearance and vital capacity that affect on athletic achivement.Physical measurement considered as one of the important factors in practicing any physical activity which demands physical features that differs from the other activity ,and the grogress with handball activity demands characterizing with some abilities and specification such as physical measurements and functional variables (Vital Capacity) which effect on each other. And ,them its influence upon sport performance level ,from this point importance is rised on the study of some physical measurement and its effect on Vital Capacity through analysing its trajectories that is to say the direct and indirect effect for these measurements on Vital Capacity by using the rule of incline (trajectory analysis ) and it is an applied study . The purpose from it is to know which measurements is more effective on vital capacity.

In the event that the firm produces many products, the number of constraints and bottlenecks multiple to require the work to manage these constraints effectively and efficiently and directing the firm resources towards the production mix of the products which maximize profit, a model of linear programming models is used to achieve optimum production mix which maximizes profits therefore it can be considered as one of the tools of the theory of constraints.

Texture extraction and analysis are regarded as important operations in image processing field for various computer applications.
In this research, fractional dimension has been used in the process of texture extraction and analysis of various medical images. 2D Variation and Box Counting algorithms were used to extract the textured image and segment the radiation image based on the extracted textures.
This research first adopted common and used methods in 2D variation algorithm, as well as a proposed method to adopt this algorithm. Practical experiments showed the efficiency of the proposed method in the process of texture extraction with execution time less than the time required by other common used methods.

The main of this study is to give model to estimate Websites quality by Artificial Neural Network and fuzzy system to arrange these websites according to there quality depending on the values that get it during the application of the model on these webs, this models contain the most important element suggested for quality criteria like the content quality, availability, reliability, performances and some other elements that must took care when estimating websites with possibility to apply it on all webs by using numeral numbers to meager these categories, in order to get a syntactic model can depended on in estimating websites for easy access to good websites and notarization information can use it in scientific research