IRAQI JOURNAL OF STATISTICAL SCIENCES

This paper presents fits for neural network model , and comparative resulting forecasts with those obtained from Box-Jenkins Method. We use time series data of Tigris's monthly flow into Mosul city from 1950-1995. To perform a comparative . forecasting work through the Box-Jenkins and neural network doesn't mean working with two different or competing aspect ; on the contrary choosing a proper architecture of neural net works requires using the skills of statistical modeling . As for application , Box-Jenkins Method has given more appropriate forecasts than those given by feed forward artificial neural network . We used Minitab and SPSS programs in the statistical aspect and Alyuda program in the neural network aspect.

In this paper, we have proposed a modified QN-algorithm for solving a self-scaling large scale unconstrained optimization problems based on a new QN-update. The performance of the proposed algorithm is better than that used by Wei, Li, Yuan algorithm. Our numerical tests show that the new proposed algorithm seems to converge faster as compared with a standard similar algorithm in many situations .

In this paper we improve the Karmarkar's method for linear programming by using the vector of initial point with all iteration, and when , we see that the Karmaker's method can be reduced to a direct method without iteration and grantee the optimal solution. Finally the new method have been compared with Karmaker's. The numerical results show that the new method is better and faster.

Bayesian decision theory is a fundamental statistical approach to the problem of classification as for pattern recognition. It makes the assumption that the decision problem is posed in probabilistic term, and all of the relevant probability values are known. To minimize the error probability in classification problem, one must choose the state of nature that maximizes the posterior probability. Bayes formula allows us to calculate such probabilities given the prior probabilities, and the conditional densities for different classes or categories.
Statistical classification is one of the most useful analysis tool which can be used for analyzing several kinds of data in various sciences. Its interesting in separating more than one class or category from each other, when their behaviors are near each other. The Bayesian surface decision equation is one of the classification rule that produces a linear separable equation for these interacted classes, specially where they contain vectors of random variables that are identically independently distributed (iid),(or all vectors have the same distribution with same parameters values). This was done as a special case by Muller, P. & Insua, D.R, (1995), This study is a trial to generalize (Bayesian Surface Decision Equation) to produce a such (Linear Separable Equation) as a linear classifier for those classes contain random vectors distributed identically Gaussian) but with different parameters values ( mi , Σ i ), moreover, in this study the researcher tried to search for equivalence between Bayesian Surface Decision Equation, and a linear perceptron for classification for this general case, with a numeric application, (encoded data vectors) that is illustrated latter.

This research aims to finding the shortest path between two Centers S and T which are connected by a network of roads of certain length and certain material or time costs, (as the word Centre could be cities or storage depots or production centers or centers of import and export) by using the new proposed programmable algorithm which depends on converting the road network, to the matrix in which number of columns and rows is equal to the number of the arcs of the under-study network (What ever the size of the road network of the studied problem is), and its elements are: either (1) which means that there is an arc connecting the two centers, or (0) which means that there is no arc connecting the two centers, and then processing this matrix in three basic stages:
First: the stage of identifying the number of paths in the network: by counting the number of elements that are equal to (1) in the first row of the matrix, which represents the initial number of tracks, then moving to the following lines and counting the number of elements that is equal to (1) if they are more than one we will have other paths their number is equal to the initial number of elements that is equal to (1) minus one, and add this number to the number of tracks, and so on.
Second: the stage of limiting the number of arcs, which is embedded in each path by identifying the number of the line and column of the elements that are equal to (1).
Third: The stage of counting the lengths of the paths according to the arcs embedded in each path and comparing them to determine the shortest one.

Statistical inference is based generally on some estimates that are functions of the data. Resampling methods offer strategies to estimate or approximate the sampling distribution of a statistic. In this article, two resampling methods are studied, jackknife and bootstrap, where the main objective is to examine the accuracy of these methods in estimating the distribution of the regression parameters through different sample sizes and different bootstrap replications.
Keywords: Jackknife, Bootstrap, Multiple regression, Bias , Variance.

In this brief research we use graph theory, which is considered as one of the theories of Operations Research, we give a brief account of what this theory and its subsidiaries and their importance, their uses and one of the discussions dealt with the use of such theory in the field of management is to find a solution to the issue of timetables note that the timelines are not limited to the administrative side, but also intervention in many areas as well .
The issue of tables of academic lectures has been chosen because it is the closest to the perception in the university where a number of faculty members are distributed to number of rows to give lectures to these rows in order to achieve the requirement of teaching in the curriculum and during the specified period of time to achieve the condition that there is no more than one teacher to the same row at the same time and also there should be no more than a lecture to the same teacher at the same time.
After the application of graph theory and using two methods, including two color method and less tree generator method we can obtain a timetable to achieve the required conditions. These two methods are clarified by giveing a simple example was obtained timetables for appointments, although the difference in the results of both methods was very simple and both solutions were investigating conditions of the issue.

This research is dedicated to study non-stationary time series , and the ability of using fuzzy logic in order to improve forecasting. The non-stationary time series (Mixed autoregressive and moving average model) has been linked with fuzzy logic in order to get on the parameters of fuzzy time series models (Fuzzy mixed autoregressive and moving average model), and applied on monthly purchases rates data and foreign currency sales (Dollar) for the daily bid of Iraqi central bank. The fuzzy mixed autoregressive and moving average model for time series gave more appropriation forecasting than the forecasting given by fuzzy mixed autoregressive and moving average model

In this paper, we propose a number of families of estimators of ratio and regression to estimate the population variance using auxiliary information in two phase sampling, the bias and the mean square error expression for the above estimators are obtained up to terms of order only and calculate minimizing the for all estimators and also the generalization of estimators.

The operation of order estimation of a Markov chain to represent the observation chain is an important problem in realistic applications. One method of order estimation depending on intelligence technicality is represented by artificial neural networks . in this research we design an artificial neural network which is the backpropagation error (BPE) .
This research treated realistic problem in Markov chain order estimation to the weather state of raining months in Nineveh Governorate ( clearly, cloudy, rainy) , and a special algorithm has been prepared for training the network .We suggest another neural network for backpropagation error which is forecasting the weather state , which is designed depending on the past network after modify it.

In this paper many state space models are fitted. The statistical criteria used to choose the best estimated model whose parameters estimated by the Ridge regression method. This method can be used when the predicator variables are multicollinear.The data used in this paper represented by the temperature, which is the response variable and the predictor variables effulgence, wined, sunshine and evaporation.

More than three years of measurements of aerosol size-distribution and different gas and meteorological parameters made in a specific region were analyzed for this study to examine which of the meteorological and trace gas variables effect on the emergence of nucleation events. As an analysis method, we used discriminate analysis with non-parametric density estimation method. The best classification results in our data was reached with the combination of relative humidity ozone concentration and a third degree polynomial of radiation. RH appeared to have a preventing effect on the new particle formation whereas the effects of O3 and radiation were more conductive. The concentration of SO2 and NO2 also appeared to have significant effect on the emergence of nucleation events but because of the great amount of missing observation, we had to exclude them from the final analysis

This paper deals with the proplems of contamination and heteroscedasticity , by using a method (the Direct Wavelet Filter that used which some method of thresholding) and comparing it with some classical method of transformation in the complete randomized design , through a Matlab code that perform the filtering process , also SPSS was also used.

In this paper we describe single exponential smoothing method, which is used in time series forecasting, and suggest an improving to the single exponential smoothing method through adding the mean of the first differences for the time series for all predicting values of the single exponential smoothing. The improved method was compared with single exponential smoothing method by using real time series data for wheat national production for the period (1961-2002) through depending on Cumulative Forecasting Error (CFE), Mean Absolute Deviation (MAD), Mean Square Error (MSE), and Mean Absolute Percentage Error (MAPE) as criteria for comparison. It is clear that the improv method was more efficient than the single exponential smoothing method for forecasting in time series.

The need for control systems exists in many fields, such as, medicine, economics, engineering, agriculture, etc., as a result of the rapid progress in science and technology in all fields which cause a creative interaction between control theory and these fields, especially, in the field of computer science.
Other modern techniques were developed in the field of control systems such as fuzzy logic which was used in this research with discriminant function by monitoring production quality control, as a multivariable method, and comparing between fuzzy logic and discriminant function through differentiating between identical products by their materials structure entered in manufacturing method but they are different in values, and specified standards for each one of them. The fuzzy logic method overwhelmed discriminant function method as being specified for complex systems treatment to get more accurate results.

Clustering is considered as one of the most scientifical developments which the scientists reached at in the field of recent knowledge and technologies to discover the cluster's group. The clustering concept was introduced firstly by Ronald in 1955. The clustering's fundamental notion is represented in dividing the data into clusters. This research aims to using clustering for actual data modeling for the monthly salary grade of the teaching staff for one of the Mosul University's College in 2009, by using HCM algorithm to these data. Matlab software is used to write down the proposed algorithm programs. Results proved the efficiency of this algorithm in clustering the actual data and how to represent them as clusters .

In this paper, the solution of Symmetric Fuzzy Quadratic programming is considered by using fuzzy values with special Phases of Van De method , For the founded model ,we use the Kuhn Tucker conditions for solving the Quadratic programming by implementing the computer program (WINQSB) also by transforming the model to two models one of them is Linear and the other is Quadratic , of the Matlab in order to check the results. Finally, the results are very encouraging.