Search — Iraqi Digital Repository

Search results: 25 out of 856

استعمال بعض نماذج السلاسل الزمنية للتنبؤ بمحصول الرز في بعض مناطق الوسط والجنوب في العراق == Use Some Time - Series Models To Predict Rice Production In Some Middle And South Areas In Iraq For The Period (1969 - 2014

Author name: عدويه حسن يوسف الطائي

Supervisor name: غفران اسماعيل كمال

General topic: Administration and Economics

Specific topic: Applied Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

استعمال بعض الاساليب الاحصائية للتنبؤ بانتاج محصول الشلب في العراق للمدة (2025 - 2016) == Some Use Statistical Methods To Predict The Yield of Rice Production In Iraq (2016 - 2025)

Author name: عبير محمود جاسم

Supervisor name: احمد ذياب احمد العزاوي

General topic: Administration and Economics

Specific topic: Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: يعد القطاع الزراعي من القطاعات المهمة في عملية البناء الاقتصادي والاجتماعي في بلدان العالم المختلفة وان على الزراعة ان تسد حاجة السكان من الغذاء حتى يتحقق الامن الغذائي.ويعد محصول الشلب من المحاصيل الصيفية المهمة التي تزرع في المناطق المروية من وسط وجنوب | The agricultural sector of the important sectors in the process of economic and social reconstruction in the different countries of the world and on agriculture to meet the needs of the population of food so food security is achieved. The rice crop of the important summer crops grown in irrigated areas of central and southern Iraq, one of the important crops in food intake in Iraq depends upon the Iraqi individual in the food and meals. This research aims to predict the production of rice crop in Iraq for the period (from 2016 to 2025) through the time - series models using the Box and Jenkins models, which requires diagnosing the appropriate model and appropriate grade for this model to represent the phenomenon studied and based on annual data for the period analyzed (from 1975 to 2015), which was obtained from the Central Bureau of Statistics. After analyzing the data using statistical software gretl and address some of the problems that could have exposed the data and obtain the appropriate models to represent the production of rice, and in the model of the time series were tested stability of the chain in terms of both the contrast and the Mediterranean have shown results of the analysis after the trade - offs between Box and Jenkins models for series production of rice the adoption of standard Akaike Information Criterion (AIC) and the standard Schwartz Bayesian Criterion (SBC) and standard Hannan - Quinn (H - Q) to be after that model representative for the production of rice crop in Iraq is ARIMA (2,1,2) and test the independence of the wrong paradigm and then predict for the years (2016 - 2025).

استعمال التحليل العنقودي لبيان تغير نمط الانفاق في العراق للسنوات (1971 - 2010)

Author name: عبير عباس ناصر

Supervisor name: لميعة باقر جواد الجواد

General topic: Administration and Economics

Specific topic: Applied Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

تشخيص وفحص مدى الملاءمة لنماذج السلاسل الزمنية المختلطة ذات الرتب الدنيا == Identification And Diagnostic Checking For Low Order Mixed Models

Author name: عبيد محمود محسن الزوبعي

Supervisor name: عبد المجيد حمزة الناصر

General topic: Administration and Economics

Specific topic: Statistics

Degree: Doctorate

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تمثل الهدف ببحث مرحلتي التشخيص Identification وفحص مدى الملاءمة Diagnostic checking للنماذج المختلطة ARMA ذات الرتب الدنيا في مجال الزمن Time Domain وفي مجال التكرار Frequency Domain والمقارنة بين ادوات التشخيص ومعايير اختيار الرتبة واختبارات فحص مدى المل | The aim of this work is to study the two stages of identification and diagnostic checking of mixed models (ARMA) with low orders of time domain and frequency domain.The methodology of research had tackled in balanced way both the theoretical part (by using statistical theory) and experimental part (by using simulation). The thesis consisted of five chapters in addition to introduction. The first chapter contained the basic concepts of time series, stationary, mixed models with low orders and analysis tools for time domain and frequency domain.The second chapter included the identification stage as it contained the identification tools and selection criteria for models order.Also it included asggestion for two new methods of identification and anew criterion for order determination. The third chapter tackled the diagnostic checking by using a group of tests depending on time domain and frequency domain.Also it included viewing of some other aspects in time series analysis in order to open new avenues which could be traded by other researchers.The fourth chapter contained the experimental part by applying what had been depicted in previous chapters on mixed models (ARMA(1,1)), (ARMA(1,0)) and (ARMA(0,1)) and then to find out comparisions between identification tools, order selection criteria and diagnostic checking tests through giving different values of the parameters [?1,?1] and for different sizes of series and by iterating the experiments (1000) times.It had been arrived at some conclusions and recommendations which were consisting the fifth chapter

التنبؤ بحركة الطرود البريدية الصادرة والواردة باستعمال نماذج بوكس وجينكنز == Forecasting The Movement of Incoming And Outcoming Postal Parcels Using Box - Jenkins Models

Author name: عباس دوين عنكوش

Supervisor name: اسماء غالب جابر الراوي

General topic: Administration and Economics

Specific topic: Applied Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

حساسية نموذج مباراة تفاضلية == The Sensitivity of Diffential Game Model

Author name: عايدة زاوين ارشاكيان

Supervisor name: ظافر حسين رشيد النجار | لميعة باقر جواد الجواد

General topic: Administration and Economics

Specific topic: Operations Research

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

المويجات الهندسية المكيفة وتطبيقاتها في ازالة التشويش للصور الرقمية == Adaptive Geometrical Wavelets And Their Applications In Digital Image De - Noising

Author name: عامر محمد نوري المهداوي

Supervisor name: ظافر حسين رشيد النجار

General topic: Administration and Economics

Specific topic: Statistics

Degree: Doctorate

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تختلف الاهمية النسبية للمعلومات المرئية المحتواة في الصور. اذ ان البحوث الحديثة في فسلجة الرؤية اثبتت بان الدماغ البشري يتحسس لحواف مكونات الصورة بالدرجة الاولى، اما الانسجة فتاتي بالدرجة الثانية من الاهمية. دفع هذا الامر العديد من الباحثين في حقل المعالج | The importance of visual information contains in images are relatively deferent. Recent researches in psychology of vision have proven that human brain senses the edges of image objects in the first order, where image textures comes in the second order of importance. This problem had motivated researchers in the field of image processing towards finding new efficient methods that work on eliminating less important visual information which gives as a result a brief description containing the most important information. This description serves two domains : the first one is image compression, where it is possible to reduce the amount of data used in representation, which helps in reducing the required transmission time and also needs less storage size. The second one is image denoising, by reducing the importance of noisy data which helps in eliminating the noise, or at least compress its effect.Recently, it has become evident that separable transforms, such as wavelets, are not necessarily best suited for image representation due to their disability of catching line discontinuities that represent the objects edges, which make them unsuitable in giving adequate description for deferent geometrical shapes of these edges. This led to the appearance of the geometrical wavelets which has been proven its superior to nearly all of the classical wavelets in giving higher specific approximations and much more economical in data size of the image. These functions are non - separable and they are divided into two parts : adaptive functions and non - adaptive functions.This dissertation focuses on studying the first and most important function of the family of adaptive geometrical wavelets functions, the one called wedgelets, in addition to propose generalizations to it in order to make it more flexible in catching deferent curved shapes of image edges, which improves its performance in providing much more specific approximations depending on less number of coefficients. This work emphasizes basically in image denoising. And in order to increase the applicable importance of these functions, it has been suggested two methods to estimate the noise level that effects the image, and eventually choosing the best approximation according to this estimation.Finally, a comparison is made between the proposed approximation methods and the classical ones by applying on several tested images that have deferent properties, through the simulation of exposing them with deferent levels of noise. The results of this comparison shows good size of improvement in the performance achieved by the addition of these proposed functions.

استخدام مرشح الموجة الصغيرة المتقطعة في تحليل السلسلة الزمنية AR (1) ومقارنته مع مرشحات اخرى == Using Discrete Wavelets Filter In Analysis of Time Series AR(1) And Comparison With Other Filters

Author name: طه حسين علي الزبيدي

Supervisor name: عبد المجيد حمزة الناصر

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تم في هذه الاطروحة معالجة مشكلة الضوضاء (او التلوث) الذي يمكن ان تتعرض له مشاهدات السلسلة الزمنية باستخدام مرشحات الموجة الصغيرة والاعتماد على تحليل المعلومات من خلال التردد فضلا عن الزمن مقارنة مع تحليل فورير الذي يحلل المعلومات عن طريق التردد فقط مهم | This Thesis deals with the problem of Noise (or Contamination) which may encounter the time series data using Wavelet Filters and depending on the information analysis through the frequency in addition to the time compared to Fourier Analysis which analyze the information through the frequency only omitting the time factor, this was performed through the use of Discrete Wavelet Transformation as a filter to clean the data from the contamination or the noise factors by direct or with some kinds Thresholding , and then Estimate the First - order Autoregressive Model for the filtered time series observations and compare the results with what results from the use of time series observations that are contaminated and filtered by using Wiener and Kalman filters depending on some statistical Criterias, which are The Mean Square Prediction Error, Final Prediction Error, and The Mean Absolute Prediction Error.This study presents the suggested method as well, that depends on the Wavelet as Input for The Artificial Neural Network, and then use the outputs of this Network to Estimate The First - order Autoregressive Model to the time series observations that are filtered, and compare the results with the Classical Method - Neural Network, Haar Wavelet, and Daubechies Filters of the directs from second order and which used with Soft, Mid, and Hard Thresholding by depending on the statistical Criterias given before through using the simulation experiments in addition to use real data represents time series observations of sunspots, and in order to perform this analysis the researcher designed the required computer codes by using MATLAB Language.

مقارنة بعض الطرائق الحصينة في تحليل الارتباط القويم الخطي باستخدام المحاكاة مع تطبيق عملي == Comparison of Some Robust Methods In Linear Canonical Correlation Analysis of Simulation Used With Practical Application

Author name: طارق عزيز صالح

Supervisor name: لقاء علي محمد العلوي

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تبرز اهمية الحاجة الى طرائق تقدير كفء والتي تسمى بالطرائق الحصينة (Robust Methods) عندما تكون بيانات الظاهرة المدروسة ملوثة، او وجود شواذ في المشاهدات، والتي ينتج بسببها مقدرات تؤدي الى زيادة او نقصان في متوسط مربعات الخطا (MSE) مما يؤدي الى استدلال احص | The important of the necessity to my efficient estimation methods, which are called the robust methods, appears when the data of the studied phenomenon are contaminated, it means the observations contains outliers,which may produce estimators which result in increasing (decreasing) in the (MSE), That would leads to an inaccurate statistical inference. From this point was the good behind this research in reaching robust estimators of canonical correlation analysis that can be achieved through the study of some robust methods such as (estimators - M, estimators - MVE, estimators - MCD and estimators - S).for the failure of the classical methods of estimation of canonical correlation analysis at containing data on the ratio contamination (outliers), that leads to the derivation of incorrect covariance matrix consequently to correlation matrix and a series of relations between the incorrect variables. Therefore the robust methods were used in the calculate of covariance matrix which would lead to robust matrices to robust canonical correlation analysis.In order to achieve the objectives of the research, it was divided into four chapters.The first chapter included the introduction, purpose of search and review of literature, the second chapter tackled the theoretical aspect of the robust methods in canonical correlation analysis as well as some of the important concepts, the third chapters deals with the application aspect in which tow types of applied studies were made, The first one uses the simulation method to compare among the studied methods of estimation in canonical correlation analysis and detect the best of the estimator depending on the two statistical measurements bias mean and mean square error which renders the minimum MSE of canonical correlation analysis, the second study uses the truthful data to verity the performance in a practical actuality.Finally, the fourth chapter included the conclusion, recommendations to which the researcher has arrived. In general it is regarded that MCD estimator is the best in the estimation of canonical correlation analysis in comparison with the other studied methods of estimation.

التنبؤ باعداد طلبة المدارس في محافظة بغداد لغاية 2024 == Forecasting Numbers of School Students In Baghdad Province Up To 2024

Author name: صهيب اسماعيل عبد اللطيف

Supervisor name: اسماء نجم عبد الله

General topic: Administration and Economics

Specific topic: Applied Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: جميعنا يعلم ان التعليم يبدا من خلال مراحل متعددة والتي تشكل سلسلة زمنية يمكن الاستفادة منها والتنبؤ بها من اجل تحسين وتطوير هذا القطاع المهم, مما لاشك فيه ان للتنبؤ دورا مهما وبارزا في عملية اتخاذ القرارات التي ترسم مسار الادارات والهيئات لما له من نتا | We all know that education starts through multiple stages, which constitute a time series that can be tapped and predictable in order to improve and develop this important sector.There is no doubt that to predict an important and prominent role in the decision - making process that paint departments and agencies track because of its ineffective results, because it contains a range of methods and statistical methods, including the methods, time series analysis and style Box and Jenkins, one of them, which depends on the time series of the phenomenon being studied analysis to provide the so - called autoregressive moving averages models characterized that methodology that can predict the data that appear in the future, based on time series data of interest in the past.The research aims to predict the numbers of school students, and educational supplies to the province of Baghdad, using the best model among the Box _ Jenkins models.The most research data numbers, primary and secondary school students in the province of Baghdad and for the period from 1969 until 2014, equivalent to (45) Show, which build a better model was to predict the numbers of students for both phases (primary, secondary), where the researcher found a better specimen prediction for primary school students was the specimen is ARIMA (1,1,3) the best specimen for secondary students is ARIMA (1,1,0) and through a table predictive values for each of the primary and secondary enables the researcher to assess the educational supplies by developments in the preparation of students increase according to the regulations and instructions of the Ministry of Education.Search was divided into four chapters shows the first chapter introduction research and objective of the research and review Reference The second chapter took the theoretical aspect of identifying the theoretical bases of Box_cengnz models in time - series analysis of either the third quarter included the practical side.Was estimated (the number of people to primary and secondary education, the number of primary and secondary schools, the number of teachers, number of teachers, the number of supervisors, the number of supervisors specialists, the number of school trips, the number of books for primary and secondary).The fourth chapter and the latter is for the conclusions and recommendations reached by the researcher that benefit the sound development of plans for the advancement of the educational sector

مقارنة بعض طرائق تقدير المعلمة والمعولية لانموذج ريلي للفشل لبيانات تامة وبيانات تحت المراقبة من النوع الاول باستخدام المحاكاة == A Comparison of Some Estimation Methods of Parameter And Reliability of Rayleigh Failure Model For A Complete Data And Type One Censored Data By Using Simulation

Author name: صبا صباح احمد الجميلي

Supervisor name: عبد المجيد حمزة الناصر

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تعرض الكثير من الباحثين الى انموذج ويبل للفشل (Weibull Failure Model) وتطبيقاته المختلفة كونه احد نماذج الفشل الشائعة الى جانب اهميته في حقل المعولية واختبارات الحياة. فـي هـذا البحث تـم تقديـر معلمــة القياس ودالـة المعوليـة لانموذج ريلـي للفشـل (Rayle | Weibull Failure Model and its different applications have been studied by many researchers, since it is one of the well known failure models in addition to its importance in the reliability field and life tests.This research estimated the scale parameter and reliability function for Rayleigh Failure Model, which is one of the well known failure model in the reliability field and life tests and the signal analysis.The research focused on the comparison between some of the well known estimation methods (classical and Bayesian) for the scale parameter and reliability function of this model by using the Maximum likelihood, Moments, White and Standard Bayes estimation methods.The methodology of the research depends on theoretical study, the methods of classical and Bayesian estimation has been determined elaborately to arrive to the estimations forms of reliability.Also this research depends on an experimental study by designing number of simulation experiments using various values of parameters and sample sizes, this experiment replicated to get high homogeny for the comparison among the estimation methods.Two kinds of data are used in the research : 1 - Complete Data.2 - Time - Censored Data. The researcher suggested two Bayesian methods for estimation in case of complete data; the first is a loss function obtained by compressing the well known square loss function which called it (The Compressed Bayes Method), and the second is a prior function which is called (The Developed Bayes Method). The comparison between the two proposed methods is done to show the effect of Bayesian estimator, then a comparison done between the best proposed methods with the best well known methods to show which estimator is the more accurate to be used for estimation the scale parameter and reliability function for Rayleigh failure model.The results of these experiments show that White method is the best from the other methods which are used in this research in estimating the scale parameter and reliability function, and the results show that the proposed Compressed Bayes method is better than the Standard Bayes and the Maximum Likelihood methods, while the proposed Developed Bayes method verified that it is better than the Moment method in estimating the scale parameter and the reliability function; and better than the Maximum Likelihood estimation method in estimating the reliability function, and using any of the proposed methods (Compressed Bayes, Developed Bayes) is better than the Standard Bayes method. In case of censored data the simulation experiments proved that the Maximum Likelihood estimation method is better than the Standard Bayes estimation method. A comparison of preference estimation methods is done by using the two standards; Mean Square Error (MSE) and Mean Absolute Percentage Error (MAPE).

تاثير تغير اسعار النفط على موازنة العراق العامة للفترة (2003 - 2014) == The Effect of Changes of Oil Prices By The General Budget For Iraq On Duration (2003 - 2014)

Author name: شيماء محمد احمد

Supervisor name: فارس طاھر حسن الكواز

General topic: Administration and Economics

Specific topic: Applied Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

استعمال بعض نماذج السلاسل الزمنية للتنبؤ باعداد التلاميذ المقبولين في الصف الاول الابتدائي في العراق للفترة (1969 - 2012) == Use Some Time - Series Models To Predict The Numbers of Pupils Admitted To The First - Grader In Iraq For The Period (1969 - 2012)

Author name: شيماء ابراهيم خليل

Supervisor name: هاني عبد الله حسن

General topic: Administration and Economics

Specific topic: Statistics

Degree: Higher Diploma

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

مقدرات الامكان الاعظم الموزونة الحصينة ومقارنتها مع طرائق اخرى لانموذج اللوجستك مع تطبيق عملي == Robust Weighted Maximum Likelihood Estimates And Their Comparsion With Other Methods of Logistic Model With Practical Application

Author name: شرين علي حسين

Supervisor name: نزار مصطفى جواد الصراف

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: ان معظم الدراسات في موضوع انموذج انحدار اللوجستك تاخذ طابعا اكثر تقدما في عملية التحليل الاحصائي الدقيق الذي يهدف الى الحصول على مقدرات ذات مستوى عال من الكفاءة. وتاتي اهمية الحاجة الى تطبيق الطرائق الحصينة عندما تكون بيانات الظاهرة المدروسة ملوثة , | In most of studies about the logistic regression model take in it's nature more progness in the procedure of acurate statistical analysis which aim getting estimators of a high level of efficiency , and importance of it's necessity to applicate of the robust methods, appears when the data of the studied phenomenon are contaminated , it means some of the observations vareit clearly from other observations called outliers which may produce estimators untrustworthy and so the decision will be false about the problem that we are trying to study it.From this point was the goal of this research in reaching robust estimators to estimate parameters of binary response logistic regression model through study some of theis estimators (method (WMLE) , method( M ) and method (Lp) ) and that by supposing three levels of contamination (0%,10%,30%). Also we use simulation in this study to compete between methods of studied estimator for all levels of the contaminated studies.In this thesis the researcher concluded the success of the method (WMLE (w1)) in estimate parameters of binary response logistic regression model, in comparsion with studied estimate methods by depending on statistical measures : Bias, Mean Square Error for parameters and model.Also the researcher specified a chapter for applying and using method WMLE (w1) on real data of angina pectoris patients which taking from Ibn - al - Nafees hospital for the purpose of estimate parameters of binary logistic regression model.

استعمال المجموعات الضبابية ونماذج بوكس جينكنز في السلاسل الزمنية للتنبؤ ببعض نسب التلوث في مياه الشرب لمدينة بغداد == Using The Fuzzy Sets And Box - Jenkins Models In Time Series To Prediction Some Rates of Pollution In Baghdad City

Author name: سيف عدنان سلمان

Supervisor name: احلام احمد جمعة

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

توظيف الخوارزمية الوراثية في انشاء تصاميم القطاعات غير الكاملة المتزنة == Construction of Balanced Incomplete Blocks Design (B.I.B.D.) By Using Genetic Algorithms (G.A.)

Author name: سهام دنخا خوشابا

Supervisor name: كمال علوان خلف المشهداني

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

مقارنة بعض اختبارات ملاءمة نماذج السلاسل الزمنية ذات الرتب الدنيا باستخدام المحاكاة == Comparison of Some Diagnostic Checking Tests For Time Series Models With Low Order By Using Simulation

Author name: سهاد احمد احمد

Supervisor name: احلام احمد جمعة

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: ان دراسة السلسلة الزمنية تتطلب دقة في وصف ملامح الظاهرة (العملية) التي تتولد منها، سواء من حيث تفسير طبيعة تلك الظاهرة وسلوكها واستخدام النتائج للتنبوء بسلوكها في المستقبل اخذين بنظرالعناية التحقق عما يمكن حدوثه عند تغيير بعض معلمات الانموذج. ويمكن استخ | Studying time series requires an accurate description for the features of phenomena (procedure) that generates time series according to its nature and behavior interpretation. After that, we can use these results for forecasting to the future taking under consideration : examine to what would happen after changing some of the model parameters, use the standard (depended) methods for testing the accuracy of the identified model which has ability for taking errors in the procedure, and provide balancing condition between model diagnosing and simplicity as well as. This research aims to use some of tests concern with diagnosing of the identified model, and make comparisons among these tests meanwhile. The researcher took some of tests, especially, the modern ones, represented by Dm test which depends on partial autocorrelation function, and which depends on autocorrelation function. We use ARMA models with lower order : [AR (1), MA (1), ARMA (1,1)] that Gaussian distribution. Furthermore, we use the tests concern with linear hypothesis diagnosing for the previous three models, represented by squaring the model errors. We use also, the modified tests in this field. In addition, we use some of nonlinear models, as "Threshold" model, and taking self - exciting threshold autoregressive (SETAR) as a special case. The analysis of time series considered time range under model stationary. Methodology of the research depends on both the theoretical part (Statistical theory usage), and the empirical part (Simulation). The structure of the research has mainly based on four chapters as follow : chapter ONE, consists an introduction, the goal of the research and the historical literature. While chapter TWO, covers the statistical methods that take part describing and modeling the time series, as well as explain the main concepts of time series such as "Stationary". We show also, the "mixed" model with lower order, crossing to examining the diagnosing through some of the standard and modern tests which depends on autocorrelation and partial autocorrelation functions of the standardized residual series that obtained from the original series of the model, by testing the randomization of the model errors and make checking for its convenience and accuracy. Furthermore, we use the modified tests : ( , , )and (Kwan & Sim) test represented by ( , , ) to determine the linear hypothesis diagnosing by compute the standardized residual series squares of the error in the previous three models that mentioned above. We use nonlinear model (SETAR) as well as. Chapter THREE, devoted for the empirical part which contains illustrations to all of what mentioned in chapter two. The researcher has designed some of experiments for purpose of comparing the (standard, modern and modified) tests through computing (P - value) for each model at (0.01) significant level at many simulated parameters, with different sample sizes, and for a number of periods of lags at each sample size, by replicating the experiment (1000) times. Finally, chapter FOUR comprises the conclusions and suggestions that the researcher had recommended

استعمال اشجار الانحدار التصنيفية والانحدار اللوجستي في تقدير انموذج تجميعي والمقارنة بينهما مع تطبيق عملي == Using Classification Regression Trees And Logistic Regression To Estimate Additive Model Comparison With Application

Author name: سهاد احمد احمد

Supervisor name: عمر عبد المحسن علي القيسي

General topic: Administration and Economics

Specific topic: Statistics

Degree: Doctorate

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تعاني بعض الظواهر من وجود حالة اضطراب في بياناتها بالاضافة الى صعوبة صياغتها وخصوصا مع وجود عدم وضوح في استجابتها، او لكثرة الاختلافات الجوهرية التي تعتري الوحدات التجريبية التي تم اخذ هذه البيانات منها. فعندما تزداد المجاميع لمجتمع ما, لا يمكن اعتبار ذل | Often suffer phenomena and there is disturbances in their data as well as the difficulty of formulation, especially with a lack of clarity in the response, or the large number of essential differences plaguing the experimental units that have been taking this data from them. When growing groups of a population, it can't be regarded as a homogeneous population is increasing because the dispersion between the experimental units in response to that population as well as the error will also be increasingly difficult to build a model of that population.So there was a need to include an estimation method on the classification or underlying discrimination for these pilot units using discrimination or create segments for each item of these pilot units in the hope of controlling their responses and make it more in line style.Has been resorting to the use of the method of a modern and flexible classification is based on the neighborhoods between the views represented by the response variable by classification regression trees : CART. It was the use bayesian approach and the promise of features indicative of regression analysis model classification tree to take advantage of the above information by, and collect trees for explanatory variables are all together and at every stage fork on the other hand, represented the bayesian classification regression trees : BART.In view of the development in the field of computers and taking the principle of the integration of science it has been found that modern algorithms used in the field of Computer Science swarm of birds and other originally illustrated algorithm for the purposes of technology pertaining to distinguish between images or vibrational radio, etc., can be harnessed to serve the knowledge of statistics, which is based Turn on the information production and to achieve success in it. So is the use of genetic algorithm and submit a proposal by modified genetic algorithm, in addition to C4.5 and ID3 algorithms. And without losing the generality has been used as a tool of logistic analysis as important tool.It was a whole performance along with an estimate of the general additive model method of classification of all the methods mentioned in the above. Fitted model as modern and flexible overcomes curse of dimensionality that may occur most modern phenomena of the times when their data analysis.It was the use of simulated samples of different sizes (200, 400, 600) with (1000) replicates. The application was performed on patients with diabetes data for those aged (15) years and below are taken from the sample size (200) was withdrawn from the children hospital/ Al - Eskan/ Baghdad.As proof of the goodness of classification has been using the misclassification error, and as proof of the goodness of estimation has been using the root mean squares error.Was reached important conclusions through a process of differentiation among the six methods above, the superiority of the Classification and Regression Trees Algorithm (CART) is the best in terms of the results that has been reached in empirical side, when proposed Genetic Algorithm is the best in terms of the results that has been reached in applied side.

دراسة احصائية لوفيات الاطفال الرضع لمحافظة نينوى للفترة 1987 - 2004

Author name: سما سعدي علي الهاشمي

Supervisor name: عبد الحسين زيني

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

الارتباط القويم والشبكات العصبية الاصطناعية : دراسة تطبيقية == Canonical Correlation And Artificial Neural Network (Application Study)

Author name: سليمة حمادي جاسم

Supervisor name: رفعت لازم مشعل الخميسي | عدنان شهاب حمد

General topic: Administration and Economics

Specific topic: Operations Research

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

اشتراك توزيعي ويبل وتكوين توزيع الباي ويبل == The Participation of Two Weibull Distributions And The Formation of Bi - Weibull Distribution

Author name: سلام جاسم محمد الساعدي

Supervisor name: علي عبد الحسين صالح الوكيل

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: ان الاهتمام الواسع والمتزايد بدراسة موضوع المعولية باعتباره علما يتعامل مع اعمار المعدات ولاسيما احتمالات البقاء ومتوسط الحياة وهذا يعود بالاساس الى التطور التكنولوجي السريع واستخدام الانظمة الالكترونية المعقدة في مختلف المجالات. وعلى هذا الاساس فان در | The wide increasing concern of studying the topic of reliability as a science dealing with equipment ages, especially the survival and life average probabilities rendered to the fast technological development and the utility of complicated electronic systems in various fields. On that basis, the studying the topic of reliability and the connection between the theoretical and applied sides together have a great importance because they are considered the indicator towards stating the extent of efficiency and capacity of the machine and the work system without damages for a long period of time, accordingly, the study of (Weibull Distribution) is a sample of failure functions because this distribution is suitable when failure rates are high relatively at the operating start, then these rates decrease gradually with time increase. Accordingly, the thesis includes mainly a display of a new sample of joint distributions which is known as (Bi - Weibull). This kind of distribution gives more elasticity and more accurate results.Hence, the basic statistical measures and the parameter estimation of (4 - Parameter Bi - Weibull Distribution) were found through (MLE) by using the simulation method (Inverse Formula). In order to reach the research aim, the research is divided into five chapters. Chapter one deals with the topic general backgrounds and function definition including the reliability function. Whereas chapter two is dedicated to expose the methods of estimating the parameters of the (2 - Parameter Weibull Distribution) with displaying the basic statistical measures of (2 and 3 - Parameter Weibull Distributions). As to chapter three, it includes the theoretical side of (Bi - Weibull Distribution) with the exposure of the following three states of this distribution : - (4 - Parameter Bi - Weibull Distribution) - (5 - Parameter Bi - Weibull Distribution) - (6 - Parameter Bi - Weibull Distribution) with their basic statistical measures.Chapter four manifests the research programs and simulation results as well as result analysis. Chapter five includes the conclusions and recommendations the researcher has arrived at

مقارنة بعض طرائق تقدير المعلمة التمهيدية لدوال اللب متعدد المتغيرات وتوظيفها في الدوال التمييزية مع تطبيق عملي == Comparison of Some Smoothing of Estimate Methods For Multivariate Kernel Functions And Employs Them In Discriminants Functions With Practical Application

Author name: سكينة شامل جاسم

Supervisor name: لقاء علي محمد العلوي

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: ان التحليل التمييزي (Discriminant Analysis ) غالبا ما يعاني من مشكلة صغر حجم العينة ( SSS ) Small Simple Size خاصة عندما يتم تطبيقها لتصنيف انماط عالية الابعاد مثل (التعرف على الوجه، حركة اليد وحركة الجينات الوراثية و...)، او عندما تكون البيانات لديها | The linear discriminant analysis often suffer from small sample size (sss) problem.especially when they are applied for the classification of high_dimenssional patterns. Such as (face recognition) or when data is distributed normal distribution but when they are non _ linear and has spread widely. We are dealing with this problem by using a technique of kernel discriminant analysis. because this technique depends on the kernel density estimation which belong to the class of estimates. it is commonly perform Avery important technique for visualizing data distribution and smoothing. It is unimportant to procedure in the explanatory data analysis, as that used in kernel discriminate analysis well known by researchers. To identify statistical. patterns. it is also known that these estimates are. Based mainly on the choice of bandwidth parameter end controlled on smoothing of estimation and to choice kernel function. has been the use of carton methods to estimate band width parameter it is : plugin methad Least square Cross validation method Smoothed Cross validation method Are selected as a parameter of bandwidth and use in kernel density estimation (KDE( and then employ them in kernel discriminant analysis (KDA) approach. Through the joint base for groups is a kernel discriminant rule (KDR) which rely heavily on density classified (f ?_j) and probability of prior (? ?_j).therefore the problem of multiple class(The use of several variables) will be more important and visible to be parameter of bandwidth is different when compared with various densities classified. Finally, it was real data from the general blood cancer disease and style simulation application, as it was reached that way (KDA - SCV) is best when using the above standard compared with other methods because it gave the misclassification rate least.in the practical side results showed in such a way that nine people were the first (n1... n9) of the first group are infected with disease among 50 people is infected and 12 people from the second group infected with the Lisu people living with and they (n45... n50 and n1... n6 ).

مقارنة بعض الطرائق اللبية في تقدير نماذج الانحدار اللامعلمي بوجود بيانات تامة وغير تامة == A Comparison of Some Estimation Methods For Kernel Models In Complete And Incomplete Data

Author name: سعد كاظم حمزة

Supervisor name: ظافر حسين رشيد النجار | مناف يوسف

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: ان استخدام النماذج المعلمية يتطلب العديد من الشروط الاولية التي يجب توافرها لكي تكون قراءة هذه النماذج قراءة صحيحة كما تتطلب وجود بيانات من النوع الكمي الامر الذي دفع الباحثون الى البحث عن نماذج اقل صرامة من النماذج المعلمية وتمثلت هذه النماذج بالنماذج ال | Using of parametric models require a number of preliminary conditions that should be available to make the reading of these models right as well as the presence of quantitative data which motivated the researchers to search for models with lesser terms than the parametric models represented by nonparametric models. One of nonparametric methods important for estimating nonparametric regression function known as (kernel estimation) to be used for estimating any statistical function which is (smooth estimator) free of disorders modifying the observations and approximating estimated regression function to real nonparametric regression function. Hence, the researcher showed some kernel methods to estimate the nonparametric regression function in both cases complete data (FNW, FLLS, VNW, VLLS) and incomplete data (FSNW, VSNW, FINW, VINW, FSLLS, VSLLS, FILLS, VILLS) and compared them through the simulation method as well as using different models and different sizes of samples and variants. From noticing the simulation results, it was shown that the best smoother was (FLLS) by using the first model, while by using the second model, it was shown that the best smoother was (VLLS) and in the third model was (VNW). In case of incomplete data, the best smoother was (VSLLS) by using the first and second models while by using the third model, the best smoother was (VSNW)

بعض المقدرات الحصينة لقدرة الطيف وفق الانموذج المختلط ARMA مع تطبيق عملي == Some Robust Estimators For Power Spectrum According To The Arma Models With Application

Author name: سحر طارق محمود الرحيم

Supervisor name: عبد المجيد حمزة الناصر

General topic: Administration and Economics

Specific topic: Statistics

Degree: Doctorate

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تهدف الاطروحة ايجاد افضل مقدر(التقدير الامثل) لدرجة ارتداد الانموذج المختلط ARMA ذي الدرجات المتدنية لقدرة الطيف Power Spectrum والحصول على امثلية التنبؤ، اذ ان مشكلة الاطروحة تكمن في خطا اعتماد اساليب تقدير درجة الارتداد قبل اطفاء اثر الشواذ التجميعية ( | The Aim of The thesis is finding the best estimator (optimal estimate) the degree of rebound mixed model ARMA with low scores to Power Spectrum and get optimizing forecasting, As the problem of the thesis lies the mistake of adopting methods of estimating rebound degree before extinguishing (AO) Outliers (additives Values) in an estimate operation based on data series without getting rid of the effect of the presence of Outliers. According to This thesis addressing the Existence of the Outliers, in particular, (AO) Outliers after addressing the existence of Outliers, then starting estimate the optimization for a rebound.When interest in Robust estimates we must identify the time series contaminated Outliers with a certain probability models is defined due to lack of stability in addition to the difficulty of knowing specimen probabilistic it appears at first paragraph leap suddenly are called stereotypes Bahawaz Innovation Outlier, late to be the behavior of independent gay behavior values the other called Bahawaz aggregate Additive Outliers.When interest in estimates fortified must identify the time series contaminated Outliers with a certain probability models Unlimited Parameters is defined due to lack of Stationarity in addition to the difficulty of knowing probabilistic model ,it appears at first paragraph sudden leap are called Innovation Outlier, or be an independent behavior of other values called Additive Outliers.Shows the experimental side of the Thiess through a standard MSE, MAPE preference of Robust method on the Unrobes for all sample sizes and for all primary parameters assumed values, as well as for all default frequencies values (Wi), have an adverse impact in terms of efficiency when models ARMA (0,1), ARMA ( 0.2), and very slight variation when the model ARMA (1,1) the impact of outliers and the supposed frequencies (w) = [1/4, 1 / 3.3, 1 / 2.5, ½] small.The nature of the electrical signal surveyed of heart (ECG - Electrocardiogram) a vital signal , unstable and the form of Signal spectrum reflects the Performance of the human heart muscle and its health.is random changes sharp and irregular in time series represents the work of the heart of a natural person (without straining) include what is known Outliers caused by its actual nature not the result of errors in the measurement or the register, which makes them do not follow a particular distribution, led to the difficulty of conducting Assessments for Power spectrum according to the traditional methods of estimation that assumes distributed statistically,In addition to that the data generation processes require the availability of basic assumptions is inconsistent with the nature of the Heart signal ECG, and to minimize or reduce these assumptions adopted by the simulation approach often leads to unrealistic and far from the real system results and establishes a breakthrough for good properties of these estimates that we have made it necessary to adopt the idea of Robust estimates (M - estimate) in accordance with the Robust functions (Huber, Hample, Tukey, Andrew) in addition to the method of maximum likelihood Mle according to restricted simulations in the practical side, which serves as a purification process or soften the Estimators resulting from follow of statistical estimation methods on real data to determine the best estimators of Power spectrum and the search for the best Model and section distribution , and then apply the result of a trade - offs in according to indicator the indirect transition of the speed of the signal (?/S2) between the Estimation functions and Distributions on eight of combinations of Mixed model ARMA (p, q) and estimating MSE, RMSE and selecting the optimum combination which is ARMA (2,2) with Extreem value distribution and with Hampel function at section 200, which amounted to (MSE = 0.000109).

الاتساق الذاتي وتحليل المكونات الرئيسة == The Self - Consistency And Principal Component'S Analysis

Author name: سحر طارق محمود الرحيم

Supervisor name: رفعت لازم مشعل الخميسي

General topic: Administration and Economics

Specific topic: Statistics

Degree: Master

University: University of Baghdad - Faculty Of Administration And Economics - Department Of Statistics

Language: Arabic

University location: Baghdad

First pages:

Abstract: تعني عبارة " الاتساق - الذاتي " اختبارا لمحور المكونات الرئيسة عندما يرتكز التوزيع على المحور الذي يعطي تقريبا بسيطا للتوزيع الاصلي متعدد المتغيرات.جرت عملية الاختبار بفحص نقص المطابقة Lack of Fit Test في ضوء طريقة البوتستراب (Bootstrap ) اي ( تكرار ا | The term " Self - Consistency " means ; examine the Self - consistency of a principal components axis : when a distribution is centered on a principal component axis, which provides a simple straight line approximation to a multivariate distribution.A principal component axis of a random vector X; is self - consistent if each point on the axis corresponds to the mean of X given that X projects orthogonally on to that point.this examination is done by " Bootstrap - Lack of fit " test applied on to two kinds of data set (real and simulation) with different samples (small, medium, large),depending on the probability values for getting levels of significance (to assess the self - consistency of principal component axis) which is based on the frame work of the data set, whether it will be true (towards to the elliptical form for the multivariate normal distribution) or generated data set using the restrictive simulation ( towards to the asymptotic spherical form for multivariate normal distribution ) that's for estimating the self - consistency by the hypothesis ;Hk : Yk is self - consistent for Xprincipal component analysis for factor analysis is use with the aim of analyzing the large number of variables and their effective factors by the factor rotation procedure which is assumes also the multivariate normal distribution, with proposed technique, based on the smallest column on X for different sizes of samples ( small " 5 - 20 " , medium " 25 - 50 ", and large " 60 - 100 ").High level of fit, is the result of " Bootstrap - Lack of fit" test in all sample size for all research variable's; with little difference and the best for male group of the results for female group on the size ( 5 - 35 ) , conversely for another size's.and also we saw highest level of self - consistency in the small sample size ( 5 - 20 ) for true data set using direct oblimin rotation method for male group, in the medium sample size ( 25 - 50) using the same method with direct method( None) for female group.In generated data set by normal restricted simulation technique, also high level of fit is the result of " Bootstrap - Lack of fit " test in all sample size's and for all sample Size's (small, medium, large) and for tow group's (male, female), which indicate to the usefulness of using the restricted simulation technique to take pure data set, and the model with high level of stationary.also we saw highest level of self - consistency in the different sample size's, using the varimax rotation method for tow Group's (male, female), the two method's ( none, direct oblimin rotation) in some of sample size's which refers to corresponding the normal stat with varimax rotation method ,that is well known of its object for simplified explaining the factor matrix by maximization the loading variance