التنبؤ باتجاهات أعداد المقبولين في كلية العلوم الإدارية بجامعة عدن باستخدام منهجية (Box-Jenkins)‎

Abstract EN

The research is aiming at analyzing the time series by using (Box- Jenkins) patterns through different anhytic stages (diagnosis, estimation, diagnosing scruting or testing the pattern accuracy, prediction) to find out the prediction pattern of accepted number trends in the faculty of administrative sciences at university of Aden depending on annual data between(١٩٩٥-٢٠١٤) .and through estimating the coefficient of partial and subjective correlations ,and using unit root testing Dickey and Fuller and the Augmented unit root Dickey – Fuller ,it was perceived that the series of the trends of annual accepted numbers were nonstationary , concerning unit root testing we find the time series of students ,numbers of the faculty of administrative science between (١٩٩٥-٢٠١٤) is nonstationary which indicate seasonal effets in addition to the effects of the general trend ,and to be statioriary the initial differences weretaken therefore the time series of students ,numbers between (١٩٩٥-٢٠١٤) becomes stat ionary after taking the initial differences .We came to aconclusion that the appropriate pattern for this series is the pattern ARIMA(٤٬١٬٤) is the best pattern ,because the norms Akaike and Schwarz are achieving the minimum value and the determining coefficient R-sqared is at high value ٠٫٨٠ .

Arriving at the best pattern which is ARIMA (٤٬١٬٤) Y =-٣٣٫٣٠٧٨٧ – ٠٫٣٠٤٧٣٤Yt-٤ +٠٫٩٢٣٠٩١Et-٤ Consequently, there is atransition to the prediction stahe of students numbers in the faculty of administrative sciences until the year ٢٠٢٥ toward the illustrated table (٤).