دراسة نماذج الانحدار الذاتي المشروطة بعدم تجانس التباين

الموضوعات

إحصاء

روابط وملفات

التعريف والنوع

رقم الوثيقة

fda134d0-8000-4116-8fb6-fba4b32fe647

رقم العقد

0

نوع الوسائط

Crawler

نوع المحتوى

الرسائل العلمية

صيغة المصدر

رسائل دكتوراة

نوع الملف

pdf text

أسماء الملفات

1957503_1.pdf

بيانات النشر

ترجمة العنوان

Study of Autoregressive Conditional Heteroscedasticity Models

ألقاب المؤلفين

[{"name_ar":"حســام صبحــان","title_ar":"اعداد","title_en":"Preparation"},{"name_ar":"محمد طاهر عنان","title_ar":"اشراف","title_en":"Supervision"}]

اللغة

Arabic

المصدر والدورية

اسم المصدر

دراسة نماذج الانحدار الذاتي المشروطة بعدم تجانس التباين

المحتوى والصفحات

عدد الصفحات

0

ترجمة الملخص

The thesis includes four chapters dealing with the study of auto regression models of conditional variance, the issue of estimating the of the GARCH(1,1) model was studied using the genetic algorithm and its effect on the accuracy of the prediction of the product and then proposing a new method to improve the prediction of GARCH(1,1) model. We also studied the effect of the simple exponential smoothing process on the accuracy of the GARCH prediction.CHAPTER IBASIC CONCEPTS IN PREDICTION, VOLATILITY AND TIME SERIESIn this chapter, basic concepts were presented in the construction of prediction model in the view point of statistical learning and the methodology of maximum likelihood, in addition to presenting the important standard prediction accuracy. We also discussed the concept of volatility and the ways we model it briefly, as well as key concepts in financial returns. In this chapter we alsopresent the concept of time series and classification methods and the basic compounds and their effects and the definition of stationary of the second order. Box-Jenkins’s methodology was presented in the prediction of time series in brief.CHAPTER IIAUTO REGRESSION CONDITIONAL HETEROSCEDASTIC MODELSIn this chapter we present the conditional variance models, we first addressed the concept of heteroscedasticity and the difference between it and homoscedasticity. In addition to the effect of the use of conditional distribution in the expectation interval, which shows the importance of using the conditional variance model. We also presented ARCH model in which we began to formulate the ARCH(1) model and present its most important characteristics, and then anauto regression model with error of type ARCH and review the most important characteristics of the process𝑌𝑡=𝜇+𝜑𝑌𝑡−1+𝜖𝑡.A definition of ARCH(q) also displayed and the generalized model ARCH(q) which symbolizes by GARCH. In addition to the condition of stationary of the model and the method of maximum likelihood in estimating the model GARCH. We also noted a relationship between ARMA and GARCH that GARCH model have characteristics and properties that are similar to properties of ARMAin some cases, plus if it is 𝜖𝑡~GARCH(p,q) then it will be 𝜖𝑡2~ARMA (r,p). At the end of the chapter, the theoretical predictions of GARCH models were presented in general, where theoptimal prediction 𝜖𝑡~GARCH(p,q)is zero for an infinite number of past values of the process and this problem has made GARCH model not dependent on the prediction of values of the process itself, but predicting its squares. Optimal prediction for infinite past values 𝜖𝑡is 𝜎𝑡2.CHAPTER IIITHE USE OF THE GENETIC ALGORITHM IN IMPROVING THE FORECASTS OF THE GARCH(1,1) MODEL.In this chapter we present some types of evolutionary algorithms and then present the basic concept of the genetic algorithm and its components and the steps of the standard genetic algorithm. In addition, we studied the effect of genetic operators on the forecast accuracy of the GARCH(1,1) model, so we found a significant correlation. We therefore suggested intervals of probability of crossover and mutation that produced optimal predictive values for GARCH(1,1). We also proposed a method to improve the prediction of the GARCH(1,1) model through the restimation of the model using the genetic algorithm and thus obtained better prediction accuracy of the model compared to the classical method.CHAPTER VSTUDY THE EFFECT OF EXPONENTIAL SMOOTHING ON THE GARCH MODELIn this chapter we present the principle of the method of the method of exponential smoothing, and then introduced the standard exponential smoothing. We then studied the effect of exponential smoothing on the prediction accuracy of GARCH model by studying the correlation between the values of the exponential smoothing coefficient from additive error component type taken from the classifications proposed by the two scholars Hyndman & Taylor on the accuracy of the GARCH(p,q) prediction ;p, q = 1,2.

إشراف وإعداد

الإشراف

محمد طاهر عنان

الإعداد

حســام صبحــان

الاقتباسات الببليوغرافية