استعمال اساليب السلاسل الزمنية للتنبؤ بالتضخم الشهري في العراق للسنوات 2017 - 2021 == Using Time Series Models To Forecasting The Monthly Inflation In Iraq For (2017 - 2021)
Author name:
خوله جعفر 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:
07T3616 - p.pdf
Abstract:
يشكل التضخم عائقا امام تنمية الدولة لما يقع عليه من اضرار اقتصادية واجتماعية لذا تحرص الدول على متابعة سلوكه احصائيا من اجل اخذ الحيطة والحذر والحيلولة دون ارتفاعه عن طريق اتخاذ القرارات السليمة ورسم السياسات الكفوءة.لذا يهدف البحث الى ايجاد افضل انموذ | Inflation forms an obstacle to the country's development because due to its influence on the economic and social damage, so the state is keen to follow up on its conduct statistically in order to take caution and prevent rising through sound decision - making and efficient fee policies.So research aims to forecast the inflation rates Through a series of inflation on the one hand and a series of indices on the other hand, for the period (Jan 2017 - Dec 2021) and reaching any strings best in the prediction by models (Box - Jenkins) and based on the indices data for the period (Jan 2009 - Dec 2015), through the modeling of the data in accordance with the methodology (Box - Jenkins) to build models starting from the diagnosis stage and the ending in the stage to predicting forecasting ,so several types of (ARIMA) models have been proposed, including choosing the most appropriate model specimen , according to the trade - off criteria (RMSE, MAPE , BIC, AIC, Schwarz, Hannan - Quinn), then used the possible greatest function in the estimation of the model specimen parameters, it has conducted tests (Ljung - Box, test my own confidence and Normal distribution of errors test) to determine the appropriateness of the model diagnosed, and then made a future forecasts according to the model, specimen It is better to get of the results based on the ready softwares (SPSS - 22) and the program (Gretl - 1.1) and the results that have been reached were, series of inflation rates stable in mean ,seires indices of consumer prices is'nt stable, where it has a general trend stabilized after taking the first difference and depending on the test (Augmented Dickey Fuller), and was reached that the best model for forecasting inflation of inflation series is a specimen ARIMA (0,0,2) (1,0,1)12 and using the indices series is a model ARIMA (0,1, 2) (1,0,0) 12 and if we make a comparison between the two models forecasting predictions concluded that the forecasting prediction by using a series of inflation gives better results than using the indices to predict a series where the average to smooth the predicted average inflation closer to the original series and therefore, the predictive values of rates inflation closer to the values of the original series viewing And that curve prediction of a series of inflation rates better than the curve predictions using the indces as he gave finer predictions and longer periods of time.
