خوارزمية التنبؤ بمعدلات التغير الموسمية والدورية مع تطبيق عملي بالحوادث المرورية في العراق

Abstract

This research aimed to time analysis concerning variations of (Seasonal, and Cyclical) according to the suggested procedure which considered that each variation are accounted a function of mixed model, in order to test an optimal of modeled fitness among statistical assumed models, as well as suggested formulas to estimates standard error for seasonal and cyclical variation rates to obtains an estimation range limits of seasonal variation's grand mean, and cyclical variation throughout creating a confidence interval for that rates. Algorithm of the suggested estimation method are presented, starting from step of accounted variation rates (Seasonal, and cyclical) components, and ending with the final step of estimation range limits of grand mean of (Seasonal, and cyclical) variation. Data analyzed, and are expressed from annuls statistical reports of traffic accidents (Clash, Overturn, Crush, and others) which are registered and published by Ministry of Planning for the time periods (2002-2014), and whipped the process of building a forecasting model of general long term trend at the application side for the time series of the total of that accidents throughout searching for more fitted of variation rates (Seasonal, and Cyclical).Represents significant of the suggested formulas of estimating standard error of seasonal, and cyclical variations for obtaining range limits of seasonal variation's grand mean, and cyclical variation's grand mean, throughout building a confidence interval for that rates, and that should be applicable after proved the normal distribution assumption for each one. As well as the similarity behavior of seasonal, and cyclical variation's rates throughout cube polynomial model of order one lagged time.